@juspay/neurolink 13.1.1 → 13.3.0

package/docs-site/static/search-index.json

@@ -115,9 +115,9 @@
 {"objectID":"2127ec4796813559375c700e37b4944f77c72f22a0410fe93bd6f5f7cff8a4c0","title":"17. Performance Benchmarks (Expected)","url":"/docs/WORKFLOW-ENGINE-LLD#17-performance-benchmarks-expected","content":"| Workflow      | Models | Judge | Latency (p50) | Latency (p95) | Cost Multiplier |\n| ------------- | ------ | ----- | ------------- | ------------- | --------------- |\n| consensus-3   | 3      | 1     | 3.2s          | 5.1s          | 4.2x            |\n| fast-fallback | 1-2    | 0     | 1.1s          | 2.8s          | 1.3x            |\n| quality-max   | 2      | 1     | 3.5s          | 4.9s          | 3.1x            |\n| multi-judge-5 | 3      | 2     | 4.8s          | 6.7s          | 5.3x            |","hierarchy":{"lvl0":"WORKFLOW ENGINE LLD","lvl1":"Workflow Engine - Low-Level Design","lvl2":"17. Performance Benchmarks (Expected)","lvl3":""}},
 {"objectID":"fe375b96a61c82af8b72bb5f813525686c6901fd124a47eaecb87b3475939ee2","title":"Phase 3: Workflow Chaining","url":"/docs/WORKFLOW-ENGINE-LLD#phase-3-workflow-chaining","content":"Planned: createWorkflowPipeline() is not in the current release.","hierarchy":{"lvl0":"WORKFLOW ENGINE LLD","lvl1":"Workflow Engine - Low-Level Design","lvl2":"Phase 3: Workflow Chaining","lvl3":""}},
 {"objectID":"2f54303255393f1b5c8695c6342964ce83cb786fd3ae00f535bf17c44cae16cd","title":"📝 Implementation Checklist","url":"/docs/WORKFLOW-ENGINE-LLD#-implementation-checklist","content":"[ ] Create src/workflow/ directory structure\n[ ] Implement types.ts with all interfaces\n[ ] Implement config.ts with Zod schemas\n[ ] Implement ensembleExecutor.ts\n[ ] Implement judgeScorer.ts\n[ ] Implement responseConditioner.ts\n[ ] Implement workflowRegistry.ts\n[ ] Implement workflowRunner.ts\n[ ] Create built-in workflows (consensus, fallback, quality-max)\n[ ] Add methods to NeuroLink class\n[ ] Export types from src/lib/index.ts\n[ ] Write unit tests (80% coverage target)\n[ ] Write integration tests\n[ ] Add JSDoc documentation\n[ ] Create user guide with examples\n[ ] Add CLI support (optional Phase 2)\n\nDocument Status: ✅ Ready for Implementation  \nNext Step: Code generation upon approval","hierarchy":{"lvl0":"WORKFLOW ENGINE LLD","lvl1":"Workflow Engine - Low-Level Design","lvl2":"📝 Implementation Checklist","lvl3":""}},
-{"objectID":"de72b620d608319199e3abf25b54298afc8fe5f294c73c538956a2feb63fe9b1","title":"The Nervous System Model","url":"/docs/about/nervous-system-model","content":"The Nervous System Model\n\nNeuroLink is built around a biological metaphor — not as decoration, but as a structural model that governs every architectural decision.\n\nThe Three Components\n\nNeurons — LLM Providers\n\nNeurons are where intelligence is generated. In NeuroLink, neurons are the AI providers, including: Anthropic, OpenAI, Google (AI Studio + Vertex), AWS (Bedrock + SageMaker), Azure, Mistral, LiteLLM, OpenRouter, Ollama, Hugging Face, DeepSeek, NVIDIA NIM, LM Studio, llama.cpp, OpenAI-compatible endpoints, TypeSafe Jev, Laya, XOR, SubQ, Perplexity Decisions, Cloudflare Clef and OpenAI Decisions (decision-only) — plus voice neurons (OpenAI TTS, ElevenLabs, Google TTS, Azure TTS, Whisper, Deepgram, Azure STT, Google STT), realtime neurons (OpenAI Realtime, Gemini Live), and media-generation neurons (image, video, music, avatar).\n\nEach provider is a different type of neuron — different capabilities, different costs, different latency profiles. NeuroLink's ProviderRegistry gives you access to all of them through one interface, switchable with a single line.\n\nThe Pipe — NeuroLink\n\nThe pipe is the vascular layer that carries streams between neurons and organs. This is NeuroLink itself.\n\nWhat the pipe does every time you call generate() or stream():\nContext Building — RAG retrieval, memory lookup, file processing merge into the prompt\nBudget Check — BudgetChecker validates the assembled context fits the model's window\nProvider Dispatch — ProviderRegistry routes to the correct neuron\nStream Emission — Tokens flow as an async iterable\nTool Interception — When the model calls a tool, the stream pauses, MCP tool executes, result injects, stream continues\nObservability — Every stage emits OpenTelemetry spans\n\nOrgans — Connectors\n\nOrgans are the applications that consume the pipe. They connect to the vascular layer and open a gateway — a specific way for people or systems to interact with AI.\n\nEvery application built on NeuroLink is an organ. Production organs today:\nAutomatic — Shopify operations hub: address intelligence, RTO risk scoring\nTara — Slack engineering assistant: conversational AI with MCP tool access\nYama — Code review judge: automated PR analysis and governance\n\nWhy This Model Works\n\nThe metaphor enforces good architecture:\n\nSeparation of concerns: Neurons (generation) and organs (consumption) are completely decoupled. Changing AI provider doesn't touch the application. Changing the application doesn't touch the provider.\n\nSingle flow direction: Intelligence flows one way — neuron → pipe → organ. There's no confusion about where logic lives.\n\nObservable by default: A vascular system you can't monitor is dangerous. Every stage of the pipe emits telemetry by design.\n\nComposable: Multiple organs can share the same pipe. One NeuroLink instance serves many connectors.\n\nExtending the System\n\nThe nervous system model scales in three directions:\nAdd neurons — New AI provider? Register it in ProviderRegistry.\nExtend the pipe — New capability (chunking strategy, reranker, compaction stage)? Add it to the pipeline.\nBuild organs — New application? Import NeuroLink, connect to the pipe, open your gateway.\n\nSee Pipe Architecture → for the technical implementation.","hierarchy":{"lvl0":"About","lvl1":"The Nervous System Model","lvl2":"","lvl3":""}},
+{"objectID":"de72b620d608319199e3abf25b54298afc8fe5f294c73c538956a2feb63fe9b1","title":"The Nervous System Model","url":"/docs/about/nervous-system-model","content":"The Nervous System Model\n\nNeuroLink is built around a biological metaphor — not as decoration, but as a structural model that governs every architectural decision.\n\nThe Three Components\n\nNeurons — LLM Providers\n\nNeurons are where intelligence is generated. In NeuroLink, neurons are the AI providers, including: Anthropic, OpenAI, Google (AI Studio + Vertex), AWS (Bedrock + SageMaker), Azure, Mistral, LiteLLM, OpenRouter, Ollama, Hugging Face, DeepSeek, NVIDIA NIM, LM Studio, llama.cpp, OpenAI-compatible endpoints, TypeSafe Jev, Laya, XOR, SubQ, Perplexity Decisions, Cloudflare Clef, OpenAI Decisions and ScaleDown (decision-only) — plus voice neurons (OpenAI TTS, ElevenLabs, Google TTS, Azure TTS, Whisper, Deepgram, Azure STT, Google STT), realtime neurons (OpenAI Realtime, Gemini Live), and media-generation neurons (image, video, music, avatar).\n\nEach provider is a different type of neuron — different capabilities, different costs, different latency profiles. NeuroLink's ProviderRegistry gives you access to all of them through one interface, switchable with a single line.\n\nThe Pipe — NeuroLink\n\nThe pipe is the vascular layer that carries streams between neurons and organs. This is NeuroLink itself.\n\nWhat the pipe does every time you call generate() or stream():\nContext Building — RAG retrieval, memory lookup, file processing merge into the prompt\nBudget Check — BudgetChecker validates the assembled context fits the model's window\nProvider Dispatch — ProviderRegistry routes to the correct neuron\nStream Emission — Tokens flow as an async iterable\nTool Interception — When the model calls a tool, the stream pauses, MCP tool executes, result injects, stream continues\nObservability — Every stage emits OpenTelemetry spans\n\nOrgans — Connectors\n\nOrgans are the applications that consume the pipe. They connect to the vascular layer and open a gateway — a specific way for people or systems to interact with AI.\n\nEvery application built on NeuroLink is an organ. Production organs today:\nAutomatic — Shopify operations hub: address intelligence, RTO risk scoring\nTara — Slack engineering assistant: conversational AI with MCP tool access\nYama — Code review judge: automated PR analysis and governance\n\nWhy This Model Works\n\nThe metaphor enforces good architecture:\n\nSeparation of concerns: Neurons (generation) and organs (consumption) are completely decoupled. Changing AI provider doesn't touch the application. Changing the application doesn't touch the provider.\n\nSingle flow direction: Intelligence flows one way — neuron → pipe → organ. There's no confusion about where logic lives.\n\nObservable by default: A vascular system you can't monitor is dangerous. Every stage of the pipe emits telemetry by design.\n\nComposable: Multiple organs can share the same pipe. One NeuroLink instance serves many connectors.\n\nExtending the System\n\nThe nervous system model scales in three directions:\nAdd neurons — New AI provider? Register it in ProviderRegistry.\nExtend the pipe — New capability (chunking strategy, reranker, compaction stage)? Add it to the pipeline.\nBuild organs — New application? Import NeuroLink, connect to the pipe, open your gateway.\n\nSee Pipe Architecture → for the technical implementation.","hierarchy":{"lvl0":"About","lvl1":"The Nervous System Model","lvl2":"","lvl3":""}},
 {"objectID":"27ef114b7e0cecc6ca45dbf3e9285980b691cd27d21e1912f2a27a7f309f68fe","title":"The Nervous System Model","url":"/docs/about/nervous-system-model#the-nervous-system-model","content":"NeuroLink is built around a biological metaphor — not as decoration, but as a structural model that governs every architectural decision.","hierarchy":{"lvl0":"About","lvl1":"The Nervous System Model","lvl2":"The Nervous System Model","lvl3":""}},
-{"objectID":"58e5ab0de0df2a0949ae12798cd53ac3133b2c253880bcbf0c573938376fd84f","title":"Neurons — LLM Providers","url":"/docs/about/nervous-system-model#neurons-llm-providers","content":"Neurons are where intelligence is generated. In NeuroLink, neurons are the AI providers, including: Anthropic, OpenAI, Google (AI Studio + Vertex), AWS (Bedrock + SageMaker), Azure, Mistral, LiteLLM, OpenRouter, Ollama, Hugging Face, DeepSeek, NVIDIA NIM, LM Studio, llama.cpp, OpenAI-compatible endpoints, TypeSafe Jev, Laya, XOR, SubQ, Perplexity Decisions, Cloudflare Clef and OpenAI Decisions (decision-only) — plus voice neurons (OpenAI TTS, ElevenLabs, Google TTS, Azure TTS, Whisper, Deepgram, Azure STT, Google STT), realtime neurons (OpenAI Realtime, Gemini Live), and media-generation neurons (image, video, music, avatar).\n\nEach provider is a different type of neuron — different capabilities, different costs, different latency profiles. NeuroLink's ProviderRegistry gives you access to all of them through one interface, switchable with a single line.","hierarchy":{"lvl0":"About","lvl1":"The Nervous System Model","lvl2":"Neurons — LLM Providers","lvl3":""}},
+{"objectID":"58e5ab0de0df2a0949ae12798cd53ac3133b2c253880bcbf0c573938376fd84f","title":"Neurons — LLM Providers","url":"/docs/about/nervous-system-model#neurons-llm-providers","content":"Neurons are where intelligence is generated. In NeuroLink, neurons are the AI providers, including: Anthropic, OpenAI, Google (AI Studio + Vertex), AWS (Bedrock + SageMaker), Azure, Mistral, LiteLLM, OpenRouter, Ollama, Hugging Face, DeepSeek, NVIDIA NIM, LM Studio, llama.cpp, OpenAI-compatible endpoints, TypeSafe Jev, Laya, XOR, SubQ, Perplexity Decisions, Cloudflare Clef, OpenAI Decisions and ScaleDown (decision-only) — plus voice neurons (OpenAI TTS, ElevenLabs, Google TTS, Azure TTS, Whisper, Deepgram, Azure STT, Google STT), realtime neurons (OpenAI Realtime, Gemini Live), and media-generation neurons (image, video, music, avatar).\n\nEach provider is a different type of neuron — different capabilities, different costs, different latency profiles. NeuroLink's ProviderRegistry gives you access to all of them through one interface, switchable with a single line.","hierarchy":{"lvl0":"About","lvl1":"The Nervous System Model","lvl2":"Neurons — LLM Providers","lvl3":""}},
 {"objectID":"96f94c64476fe5567b61e5374da6d1b5c5020e22654c6c082b8936ae3f72eb51","title":"The Pipe — NeuroLink","url":"/docs/about/nervous-system-model#the-pipe-neurolink","content":"The pipe is the vascular layer that carries streams between neurons and organs. This is NeuroLink itself.\n\nWhat the pipe does every time you call generate() or stream():\nContext Building — RAG retrieval, memory lookup, file processing merge into the prompt\nBudget Check — BudgetChecker validates the assembled context fits the model's window\nProvider Dispatch — ProviderRegistry routes to the correct neuron\nStream Emission — Tokens flow as an async iterable\nTool Interception — When the model calls a tool, the stream pauses, MCP tool executes, result injects, stream continues\nObservability — Every stage emits OpenTelemetry spans","hierarchy":{"lvl0":"About","lvl1":"The Nervous System Model","lvl2":"The Pipe — NeuroLink","lvl3":""}},
 {"objectID":"1c74200e8d4572e84dca981b50bb162912b4c6700ea7495d60fe286cfb16347f","title":"Organs — Connectors","url":"/docs/about/nervous-system-model#organs-connectors","content":"Organs are the applications that consume the pipe. They connect to the vascular layer and open a gateway — a specific way for people or systems to interact with AI.\n\nEvery application built on NeuroLink is an organ. Production organs today:\nAutomatic — Shopify operations hub: address intelligence, RTO risk scoring\nTara — Slack engineering assistant: conversational AI with MCP tool access\nYama — Code review judge: automated PR analysis and governance","hierarchy":{"lvl0":"About","lvl1":"The Nervous System Model","lvl2":"Organs — Connectors","lvl3":""}},
 {"objectID":"6c97c4e15e0de12e62fbb1994cc7b18328c9f2b0afa2b721d8f40f01e70d32f7","title":"Why This Model Works","url":"/docs/about/nervous-system-model#why-this-model-works","content":"The metaphor enforces good architecture:\n\nSeparation of concerns: Neurons (generation) and organs (consumption) are completely decoupled. Changing AI provider doesn't touch the application. Changing the application doesn't touch the provider.\n\nSingle flow direction: Intelligence flows one way — neuron → pipe → organ. There's no confusion about where logic lives.\n\nObservable by default: A vascular system you can't monitor is dangerous. Every stage of the pipe emits telemetry by design.\n\nComposable: Multiple organs can share the same pipe. One NeuroLink instance serves many connectors.","hierarchy":{"lvl0":"About","lvl1":"The Nervous System Model","lvl2":"Why This Model Works","lvl3":""}},
@@ -216,9 +216,10 @@
 {"objectID":"c63ea3a4c4aa6319f131d1f8178b30d7f3052ab09b5e6960e439ad76f9b8fa9b","title":"Role-Based Differentiation","url":"/docs/advanced/auth-architecture#role-based-differentiation","content":"roleLimits assigns per-role limits (highest limit wins for multi-role users)\nuserLimits assigns per-user overrides\nskipRoles bypasses rate limiting entirely for specified roles (e.g., super-admin)","hierarchy":{"lvl0":"Advanced","lvl1":"Authentication Architecture","lvl2":"Role-Based Differentiation","lvl3":""}},
 {"objectID":"1824c13c380df1f7461f4a8f7ce99478b257057b57adc5b2595a680ab923d094","title":"Adding a New Auth Provider","url":"/docs/advanced/auth-architecture#adding-a-new-auth-provider","content":"Create a provider class in src/lib/auth/providers/myProvider.ts that extends BaseAuthProvider\nImplement the abstract methods: authenticateToken(), createSession(), getSession(), refreshSession(), destroySession(), getUserSessions(), destroyAllUserSessions()\nRegister the provider in AuthProviderFactory.registerAll() with a dynamic import\nRegister metadata in AuthProviderRegistry.registerAll()\nAdd the type name to AuthProviderType union in src/lib/types/authTypes.ts\nAdd a typed config to src/lib/types/authTypes.ts and a discriminated union branch to NeuroLinkAuthConfig in src/lib/types/configTypes.ts\nAdd environment variable mappings to buildProviderConfig() in src/cli/commands/authProviders.ts\nExport the provider from src/lib/auth/index.ts\nAdd tests","hierarchy":{"lvl0":"Advanced","lvl1":"Authentication Architecture","lvl2":"Adding a New Auth Provider","lvl3":""}},
 {"objectID":"a63ad442a5c17bcb83737eb96e06af21ce0d91127807b57bf333100ea11535c9","title":"See Also","url":"/docs/advanced/auth-architecture#see-also","content":"Authentication Providers Guide -- user-facing guide with configuration examples\nFactory Pattern Architecture -- how NeuroLink uses factory + registry across the codebase\nMiddleware Architecture -- the broader middleware system","hierarchy":{"lvl0":"Advanced","lvl1":"Authentication Architecture","lvl2":"See Also","lvl3":""}},
-{"objectID":"4e93f12348468bad2ffb2c45eee31449794a11626f9ac799d5b5af33e79f742d","title":"Built-in Middleware Reference","url":"/docs/advanced/builtin-middleware","content":"Built-in Middleware Reference\n\nNeuroLink includes three built-in middleware components for common use cases: Analytics, Guardrails, and Auto-Evaluation. They ship ready to wire into your generate/stream calls.\n\nQuick Start\n\nEnable all built-in middleware with a single preset:\n\nOr enable specific middleware:\n\nAnalytics Middleware\n\nPurpose\n\nThe Analytics Middleware collects comprehensive usage metrics, timing data, and operational analytics for all AI operations. It's essential for monitoring production applications, tracking costs, and understanding usage patterns.\n\nKey Capabilities:\nToken usage tracking (input, output, total)\nResponse time measurement\nRequest success/failure tracking\nProvider and model information\nAutomatic metrics storage in response metadata\n\nConfiguration\n\nBasic Configuration:\n\nAdvanced Configuration:\n\nConditional Analytics (Production Only):\n\nCollected Metrics\n\n| Metric         | Type   | Description                        | Unit         |\n| -------------- | ------ | ---------------------------------- | ------------ |\n| requestId    | string | Unique identifier for this request | -            |\n| timestamp    | string | ISO 8601 timestamp                 | -            |\n| responseTime | number | Total request duration             | milliseconds |\n| usage.input  | number | Input tokens consumed              | tokens       |\n| usage.output | number | Output tokens generated            | tokens       |\n| usage.total  | number | Total tokens used                  | tokens       |\n\nOutput Format\n\nAnalytics data is automatically added to the response metadata:\n\nGenerate Response:\n\nAnalytics Object Structure:\n\nStream Response:\n\nFor streaming responses, analytics are available in the rawResponse:\n\nStream Analytics Structure:\n\nUse Cases\nCost Tracking:\nPerformance Monitoring:\nUsage Analytics Dashboard:\n\nIntegration with External Systems\n\nSend to Datadog:\n\nSend to Prometheus:\n\nGuardrails Middleware\n\nPurpose\n\nThe Guardrails Middleware provides comprehensive content filtering and policy enforcement to block or redact unsafe content, prevent prompt injection attacks, and maintain compliance with content policies.\n\nKey Capabilities:\nBad word filtering (configurable word list)\nAI model-based content safety evaluation\nPrecall evaluation (block unsafe prompts before they reach the LLM)\nStream and generate support\nConfigurable filtering actions (block, redact, log)\n\nConfiguration\n\nBasic Configuration:\n\nAdvanced Configuration with Model-Based Filtering:\n\nPrecall Evaluation (Block Unsafe Prompts):\n\nBuilt-in Filters\n\n| Filter Type            | Description                            | Action            | Configuration                     |\n| ---------------------- | -------------------------------------- | ----------------- | --------------------------------- |\n| Bad Words          | Block/redact specific words or phrases | Redact with *** | badWords: string[]              |\n| Model-Based        | Use AI to evaluate content safety      | Block if unsafe   | modelFilter.enabled: true       |\n| Precall Evaluation | Block unsafe prompts before LLM call   | Block request     | precallEvaluation.enabled: true |\n\nBad Word Filtering\n\nHow It Works:\n\nThe bad word filter scans both requests and responses for prohibited terms and replaces them with ***.\n\nExample:\n\nConfiguration:\n\nModel-Based Filtering\n\nHow It Works:\n\nUses a separate AI model to evaluate whether content is safe. The filter sends the content to the model with a safety evaluation prompt.\n\nSafety Evaluation Prompt:\n\nExample:\n\nConfiguration:\n\nPrecall Evaluation\n\nHow It Works:\n\nEvaluates the safety of the input prompt before it reaches the main LLM. If the prompt is deemed unsafe, the request is blocked entirely, saving costs and preventing unsafe content generation.\n\nEvaluation Process:\nUser submits a prompt\nGuardrails middleware intercepts in transformParams\nSafety evaluation model scores the prompt (0-1 scale)\nIf score = threshold, request proceeds to main LLM\n\nBlocked Response:\n\nConfiguration:\n\nStreaming Support\n\nGuardrails work seamlessly with streaming responses:\n\nStream Filtering:\nBad words are replaced with *** in each text delta\nModel-based filtering is not applied to streams (too slow)\nPrecall evaluation works for streams\n\nUse Cases\nContent Moderation for User-Generated Prompts:\nCompliance with Content Policies:\nProtecting Against Prompt Injection:\n\nAuto-Evaluation Middleware\n\nPurpose\n\nThe Auto-Evaluation Middleware automatically evaluates AI response quality using configurable criteria. It can trigger retries for low-quality responses and provide quality metrics for monitoring.\n\nKey Capabilities:\nAutomatic quality evaluation after each response\nConfigurable evaluation criteria (relevance, accuracy, coherence, etc.)\nBlocking and non-blocking modes\nIntegration with custom evaluation providers\nQuality score thresholds\n\nConfiguration\n\nBasic Configuration:\n\nAdvanced Configuration:\n\nEvaluation Criteria\n\nDefault evaluation criteria (can be customized):\n\n","hierarchy":{"lvl0":"Advanced","lvl1":"Built-in Middleware Reference","lvl2":"","lvl3":""}},
+{"objectID":"4e93f12348468bad2ffb2c45eee31449794a11626f9ac799d5b5af33e79f742d","title":"Built-in Middleware Reference","url":"/docs/advanced/builtin-middleware","content":"Built-in Middleware Reference\n\nNeuroLink includes three built-in middleware components for common use cases: Analytics, Guardrails, and Auto-Evaluation. They ship ready to wire into your generate/stream calls.\n\nQuick Start\n\nEnable all built-in middleware with a single preset:\n\nOr enable specific middleware:\n\nDefaults for every call\n\nInstead of repeating middleware on each generate() or stream() call, set it once. From lowest to highest precedence:\n\n| Layer       | How                                                  |\n| ----------- | ---------------------------------------------------- |\n| Environment | NEUROLINK_MIDDLEWARE_PRESET=default\\|all\\|security |\n| Global      | configureGlobalMiddleware({ preset: \"all\" })       |\n| Instance    | new NeuroLink({ middleware: { preset: \"all\" } })   |\n| Per call    | generate({ middleware: { preset: \"all\" } })        |\n\nmiddlewareConfig merges per middleware id; lists and other keys set by a higher layer replace the lower layer's value. With no default set, calls are unchanged. See Guardrails: Defaults and precedence for the full rules.\n\nAnalytics Middleware\n\nPurpose\n\nThe Analytics Middleware collects comprehensive usage metrics, timing data, and operational analytics for all AI operations. It's essential for monitoring production applications, tracking costs, and understanding usage patterns.\n\nKey Capabilities:\nToken usage tracking (input, output, total)\nResponse time measurement\nRequest success/failure tracking\nProvider and model information\nAutomatic metrics storage in response metadata\n\nConfiguration\n\nBasic Configuration:\n\nAdvanced Configuration:\n\nConditional Analytics (Production Only):\n\nCollected Metrics\n\n| Metric         | Type   | Description                        | Unit         |\n| -------------- | ------ | ---------------------------------- | ------------ |\n| requestId    | string | Unique identifier for this request | -            |\n| timestamp    | string | ISO 8601 timestamp                 | -            |\n| responseTime | number | Total request duration             | milliseconds |\n| usage.input  | number | Input tokens consumed              | tokens       |\n| usage.output | number | Output tokens generated            | tokens       |\n| usage.total  | number | Total tokens used                  | tokens       |\n\nOutput Format\n\nAnalytics data is automatically added to the response metadata:\n\nGenerate Response:\n\nAnalytics Object Structure:\n\nStream Response:\n\nFor streaming responses, analytics are available in the rawResponse:\n\nStream Analytics Structure:\n\nUse Cases\nCost Tracking:\nPerformance Monitoring:\nUsage Analytics Dashboard:\n\nIntegration with External Systems\n\nSend to Datadog:\n\nSend to Prometheus:\n\nGuardrails Middleware\n\nPurpose\n\nThe Guardrails Middleware provides comprehensive content filtering and policy enforcement to block or redact unsafe content, prevent prompt injection attacks, and maintain compliance with content policies.\n\nKey Capabilities:\nBad word filtering (configurable word list)\nAI model-based content safety evaluation\nPrecall evaluation (block unsafe prompts before they reach the LLM)\nStream and generate support\nConfigurable filtering actions (block, redact, log)\n\nConfiguration\n\nBasic Configuration:\n\nAdvanced Configuration with Model-Based Filtering:\n\nPrecall Evaluation (Block Unsafe Prompts):\n\nBuilt-in Filters\n\n| Filter Type            | Description                            | Action            | Configuration                     |\n| ---------------------- | -------------------------------------- | ----------------- | --------------------------------- |\n| Bad Words          | Block/redact specific words or phrases | Redact with *** | badWords: string[]              |\n| Model-Based        | Use AI to evaluate content safety      | Block if unsafe   | modelFilter.enabled: true       |\n| Precall Evaluation | Block unsafe prompts before LLM call   | Block request     | precallEvaluation.enabled: true |\n\nBad Word Filtering\n\nHow It Works:\n\nThe bad word filter scans both requests and responses for prohibited terms and replaces them with ***.\n\nExample:\n\nConfiguration:\n\nModel-Based Filtering\n\nHow It Works:\n\nUses a separate AI model to evaluate whether content is safe. The filter sends the content to the model with a safety evaluation prompt.\n\nSafety Evaluation Prompt:\n\nExample:\n\nConfiguration:\n\nPrecall Evaluation\n\nHow It Works:\n\nEvaluates the safety of the input prompt before it reaches the main LLM. If the prompt is deemed unsafe, the request is blocked entirely, saving costs and preventing unsafe content generation.\n\nEvaluation Process:\nUser submits a prompt\nGuardrails middleware intercepts in transformParams\nSafety evaluation model scores the prompt (0-1 scale)\nIf score = threshold, request proceeds to main LLM\n\nBlocked Response:\n\nConfiguration:\n\nStreaming Support\n\nGuardrails work seamlessly with streaming responses:\n\nStream Filtering:\nBad words are replaced with *** in each text delta\nModel-based filtering is not applied to streams (too","hierarchy":{"lvl0":"Advanced","lvl1":"Built-in Middleware Reference","lvl2":"","lvl3":""}},
 {"objectID":"ac9e7975939e43e8c0e52e2845fcd447f81788b86aafb183213385a1f69ee283","title":"Built-in Middleware Reference","url":"/docs/advanced/builtin-middleware#built-in-middleware-reference","content":"NeuroLink includes three built-in middleware components for common use cases: Analytics, Guardrails, and Auto-Evaluation. They ship ready to wire into your generate/stream calls.","hierarchy":{"lvl0":"Advanced","lvl1":"Built-in Middleware Reference","lvl2":"Built-in Middleware Reference","lvl3":""}},
 {"objectID":"63a19a2a4b7e313ef69e05e0d386ce69144f3ddae6c5022b20dc7dd0c9be5cd5","title":"Quick Start","url":"/docs/advanced/builtin-middleware#quick-start","content":"Enable all built-in middleware with a single preset:\n\nOr enable specific middleware:","hierarchy":{"lvl0":"Advanced","lvl1":"Built-in Middleware Reference","lvl2":"Quick Start","lvl3":""}},
+{"objectID":"a4cefe7950398abdba5bb64b133a18b0386c18ac4a6760db04a0fca5b6f99f27","title":"Defaults for every call","url":"/docs/advanced/builtin-middleware#defaults-for-every-call","content":"Instead of repeating middleware on each generate() or stream() call, set it once. From lowest to highest precedence:\n\n| Layer       | How                                                  |\n| ----------- | ---------------------------------------------------- |\n| Environment | NEUROLINK_MIDDLEWARE_PRESET=default\\|all\\|security |\n| Global      | configureGlobalMiddleware({ preset: \"all\" })       |\n| Instance    | new NeuroLink({ middleware: { preset: \"all\" } })   |\n| Per call    | generate({ middleware: { preset: \"all\" } })        |\n\nmiddlewareConfig merges per middleware id; lists and other keys set by a higher layer replace the lower layer's value. With no default set, calls are unchanged. See Guardrails: Defaults and precedence for the full rules.","hierarchy":{"lvl0":"Advanced","lvl1":"Built-in Middleware Reference","lvl2":"Defaults for every call","lvl3":""}},
 {"objectID":"ba5e28fcd0f3518c621fa4ff5d661252622d1d582be75671f9246ba89b846b94","title":"Purpose","url":"/docs/advanced/builtin-middleware#purpose","content":"The Analytics Middleware collects comprehensive usage metrics, timing data, and operational analytics for all AI operations. It's essential for monitoring production applications, tracking costs, and understanding usage patterns.\n\nKey Capabilities:\nToken usage tracking (input, output, total)\nResponse time measurement\nRequest success/failure tracking\nProvider and model information\nAutomatic metrics storage in response metadata","hierarchy":{"lvl0":"Advanced","lvl1":"Built-in Middleware Reference","lvl2":"Purpose","lvl3":""}},
 {"objectID":"f097d3aae3b2f08f80b1488048983d76d09a53d821e1249af7c8c2ef18425d68","title":"Configuration","url":"/docs/advanced/builtin-middleware#configuration","content":"Basic Configuration:\n\nAdvanced Configuration:\n\nConditional Analytics (Production Only):","hierarchy":{"lvl0":"Advanced","lvl1":"Built-in Middleware Reference","lvl2":"Configuration","lvl3":""}},
 {"objectID":"0ed403ca8e213ef3e9c955b807bced0bca556ca4e388f7006bd855c14a3181a5","title":"Collected Metrics","url":"/docs/advanced/builtin-middleware#collected-metrics","content":"| Metric         | Type   | Description                        | Unit         |\n| -------------- | ------ | ---------------------------------- | ------------ |\n| requestId    | string | Unique identifier for this request | -            |\n| timestamp    | string | ISO 8601 timestamp                 | -            |\n| responseTime | number | Total request duration             | milliseconds |\n| usage.input  | number | Input tokens consumed              | tokens       |\n| usage.output | number | Output tokens generated            | tokens       |\n| usage.total  | number | Total tokens used                  | tokens       |","hierarchy":{"lvl0":"Advanced","lvl1":"Built-in Middleware Reference","lvl2":"Collected Metrics","lvl3":""}},
@@ -325,7 +326,7 @@
 {"objectID":"83274029c906dfe570738346d3b0ee6d77dd8f4d2eb9dba30a0b6a8c25e1fb03","title":"Error Handling","url":"/docs/advanced/memory-integration#error-handling","content":"Memory is designed to never crash the host application:\nEvery public method is wrapped in try-catch\nget() returns null on error — the call continues without memory context\nadd() silently fails on error — the generate/stream result is not affected\nStorage initialization errors disable memory for that instance","hierarchy":{"lvl0":"Advanced","lvl1":"Memory Integration with Hippocampus","lvl2":"Error Handling","lvl3":""}},
 {"objectID":"3652380f1c43e08d5d18dfa0862f10a3a51f30f48679a7635ce026df2e7bc51a","title":"Production Checklist","url":"/docs/advanced/memory-integration#production-checklist","content":"[ ] Use S3 or Redis storage (not SQLite) in production\n[ ] Set HC_LOG_LEVEL=warn or higher in production\n[ ] Ensure userId is stable and unique per user across sessions\n[ ] For multi-tenant: use tenant-scoped prefixes or collection names\n[ ] Monitor Memory retrieval failed and Memory storage failed warnings in logs\n[ ] Verify the condensation LLM provider is configured and has sufficient quota","hierarchy":{"lvl0":"Advanced","lvl1":"Memory Integration with Hippocampus","lvl2":"Production Checklist","lvl3":""}},
 {"objectID":"417bcc8fbbc98c4053d46139d169f1e6f165888a5dceb6e2327b15c8bb8473e4","title":"See Also","url":"/docs/advanced/memory-integration#see-also","content":"Memory Guide - Quick start and configuration reference\nConversation Memory - Session-based conversation history\nContext Compaction - Automatic context window management","hierarchy":{"lvl0":"Advanced","lvl1":"Memory Integration with Hippocampus","lvl2":"See Also","lvl3":""}},
-{"objectID":"16f1c17ef8c078b33ae6aa26f8107300cca478f6d88a74dab2450a7762f3a54b","title":"Middleware System Architecture","url":"/docs/advanced/middleware-architecture","content":"Middleware System Architecture\n\nOverview\n\nNeuroLink's middleware system provides a powerful and flexible way to intercept, modify, and enhance AI requests and responses. Middleware enables you to implement cross-cutting concerns like authentication, logging, analytics, content filtering, and auto-evaluation without modifying your core application logic.\n\nWhy Middleware Matters:\nRequest Interception: Modify requests before they reach the AI provider\nResponse Processing: Transform, filter, or validate AI responses\nCross-Cutting Concerns: Implement authentication, logging, rate limiting, and caching in a centralized way\nComposability: Chain multiple middleware components together\nSeparation of Concerns: Keep business logic separate from infrastructure concerns\n\nKey Benefits:\nProduction-ready middleware for common use cases (analytics, guardrails, auto-evaluation)\nFactory pattern for easy middleware management\nPriority-based execution ordering\nProvider-specific conditional execution\nBuilt on NeuroLink's own wrapLanguageModel (src/lib/middleware/wrapLanguageModel.ts), which keeps the familiar wrap-a-model shape without depending on a third-party SDK\n\nArchitecture Diagram\n\nRequest Lifecycle\n\nThe middleware system processes requests through four distinct phases:\n\nPhase 1: Pre-Request (transformParams)\n\nMiddleware in this phase runs before the AI provider call, allowing you to:\nValidate input: Check request parameters for validity\nAuthenticate/Authorize: Verify user permissions\nTransform requests: Modify or enrich request parameters\nApply guardrails: Block requests with unsafe content using precall evaluation\nRate limiting: Enforce request quotas\n\nExample Use Cases:\nPrecall guardrails evaluation (blocking unsafe prompts)\nRequest parameter validation\nAdding authentication context\nModifying prompts based on user preferences\n\nPhase 2: Provider Execution\n\nThe actual AI provider call happens between middleware phases:\nRequest sent to configured provider (OpenAI, Anthropic, Vertex, etc.)\nProvider processes the request\nResponse received from provider\n\nThis phase is not middleware - it's the core AI operation that middleware wraps around.\n\nPhase 3: Post-Response (wrapGenerate/wrapStream)\n\nMiddleware in this phase runs after the AI provider responds, allowing you to:\nCollect analytics: Track token usage, response times, costs\nFilter content: Apply guardrails to block/redact unsafe responses\nEvaluate quality: Auto-evaluate response quality and trigger retries\nTransform responses: Modify or enrich the response\nCache results: Store responses for future use\n\nExample Use Cases:\nAnalytics and metrics collection\nContent filtering and safety checks\nResponse quality evaluation\nResponse caching\nLogging and auditing\n\nPhase 4: Error Handling\n\nIf an error occurs at any stage, error handling middleware can:\nLog errors: Record error details for debugging\nTransform errors: Convert provider errors to user-friendly messages\nImplement fallbacks: Retry with different providers\nAlert monitoring: Send alerts to monitoring systems\n\nExample Use Cases:\nError logging and tracking\nProvider fallback on failure\nRetry logic with exponential backoff\nUser-friendly error messages\n\nMiddleware Chain\n\nExecution Order\n\nMiddleware executes in priority order, where higher priority values run first:\n\nImportant Notes:\ntransformParams runs before wrapGenerate/wrapStream\nWithin the same priority, registration order determines execution\nMiddleware can be conditionally enabled based on provider, model, or custom logic\n\nChain Configuration\n\nConfigure which middleware to enable and their order:\n\nAvailable Presets\n\n| Preset     | Middleware Enabled     | Use Case               |\n| ---------- | ---------------------- | ---------------------- |\n| default  | Analytics only         | Basic usage tracking   |\n| all      | Analytics + Guardrails | Production with safety |\n| security | Guardrails only        | Security-focused       |\n| Custom     | Your choice            | Define your own        |\n\nFactory Pattern\n\nMiddlewareFactory Class\n\nThe MiddlewareFactory is the central component for managing middleware:\n\nCreating Middleware Instances\n\nBasic Usage:\n\nAdvanced Configuration:\n\nRegistry System\n\nRegistering Middleware\n\nThe MiddlewareRegistry manages all registered middleware:\n\nRegistration Example:\n\nDiscovering Middleware\n\nList all registered middleware:\n\nGet specific middleware:\n\nCheck if middleware is registered:\n\nMiddleware Metadata\n\nEvery middleware must provide metadata:\n\nExample:\n\nTypeScript Interfaces\n\nNeuroLinkMiddleware\n\nThe core middleware interface, which combines the model-middleware contract with NeuroLink metadata:\n\nLanguageModelMiddleware\n\nNeuroLink declares this contract itself — it no longer comes from the Vercel AI\nSDK, which is not a dependency. The shape follows the v3 model protocol:\n\nMiddlewareContext\n\nContext information passed to middleware:\n\nMiddlewareConfig\n\nConfiguration for individual middleware:\n\nMiddlewareFactoryOptions\n\nOptions for creating and ","hierarchy":{"lvl0":"Advanced","lvl1":"Middleware System Architecture","lvl2":"","lvl3":""}},
+{"objectID":"16f1c17ef8c078b33ae6aa26f8107300cca478f6d88a74dab2450a7762f3a54b","title":"Middleware System Architecture","url":"/docs/advanced/middleware-architecture","content":"Middleware System Architecture\n\nOverview\n\nNeuroLink's middleware system provides a powerful and flexible way to intercept, modify, and enhance AI requests and responses. Middleware enables you to implement cross-cutting concerns like authentication, logging, analytics, content filtering, and auto-evaluation without modifying your core application logic.\n\nWhy Middleware Matters:\nRequest Interception: Modify requests before they reach the AI provider\nResponse Processing: Transform, filter, or validate AI responses\nCross-Cutting Concerns: Implement authentication, logging, rate limiting, and caching in a centralized way\nComposability: Chain multiple middleware components together\nSeparation of Concerns: Keep business logic separate from infrastructure concerns\n\nKey Benefits:\nProduction-ready middleware for common use cases (analytics, guardrails, auto-evaluation)\nFactory pattern for easy middleware management\nPriority-based execution ordering\nProvider-specific conditional execution\nBuilt on NeuroLink's own wrapLanguageModel (src/lib/middleware/wrapLanguageModel.ts), which keeps the familiar wrap-a-model shape without depending on a third-party SDK\n\nArchitecture Diagram\n\nRequest Lifecycle\n\nThe middleware system processes requests through four distinct phases:\n\nPhase 1: Pre-Request (transformParams)\n\nMiddleware in this phase runs before the AI provider call, allowing you to:\nValidate input: Check request parameters for validity\nAuthenticate/Authorize: Verify user permissions\nTransform requests: Modify or enrich request parameters\nApply guardrails: Block requests with unsafe content using precall evaluation\nRate limiting: Enforce request quotas\n\nExample Use Cases:\nPrecall guardrails evaluation (blocking unsafe prompts)\nRequest parameter validation\nAdding authentication context\nModifying prompts based on user preferences\n\nPhase 2: Provider Execution\n\nThe actual AI provider call happens between middleware phases:\nRequest sent to configured provider (OpenAI, Anthropic, Vertex, etc.)\nProvider processes the request\nResponse received from provider\n\nThis phase is not middleware - it's the core AI operation that middleware wraps around.\n\nPhase 3: Post-Response (wrapGenerate/wrapStream)\n\nMiddleware in this phase runs after the AI provider responds, allowing you to:\nCollect analytics: Track token usage, response times, costs\nFilter content: Apply guardrails to block/redact unsafe responses\nEvaluate quality: Auto-evaluate response quality and trigger retries\nTransform responses: Modify or enrich the response\nCache results: Store responses for future use\n\nExample Use Cases:\nAnalytics and metrics collection\nContent filtering and safety checks\nResponse quality evaluation\nResponse caching\nLogging and auditing\n\nPhase 4: Error Handling\n\nIf an error occurs at any stage, error handling middleware can:\nLog errors: Record error details for debugging\nTransform errors: Convert provider errors to user-friendly messages\nImplement fallbacks: Retry with different providers\nAlert monitoring: Send alerts to monitoring systems\n\nExample Use Cases:\nError logging and tracking\nProvider fallback on failure\nRetry logic with exponential backoff\nUser-friendly error messages\n\nMiddleware Chain\n\nExecution Order\n\nMiddleware executes in priority order, where higher priority values run first:\n\nImportant Notes:\ntransformParams runs before wrapGenerate/wrapStream\nWithin the same priority, registration order determines execution\nMiddleware can be conditionally enabled based on provider, model, or custom logic\n\nChain Configuration\n\nConfigure which middleware to enable and their order:\n\nAvailable Presets\n\n| Preset     | Middleware Enabled     | Use Case               |\n| ---------- | ---------------------- | ---------------------- |\n| default  | Analytics only         | Basic usage tracking   |\n| all      | Analytics + Guardrails | Production with safety |\n| security | Guardrails only        | Security-focused       |\n| Custom     | Your choice            | Define your own        |\n\nsecurity switches guardrails on and nothing more: it ships no word list and no model filter, so it redacts nothing until badWords, modelFilter or precallEvaluation is configured. A preset, and any other MiddlewareFactoryOptions, can also be set as a default for every call: NEUROLINK_MIDDLEWARE_PRESET (a preset name only), configureGlobalMiddleware() and new NeuroLink({ middleware }), each overridden by the next and finally by the call's own middleware. See Guardrails: Defaults and precedence.\n\nFactory Pattern\n\nMiddlewareFactory Class\n\nThe MiddlewareFactory is the central component for managing middleware:\n\nCreating Middleware Instances\n\nBasic Usage:\n\nAdvanced Configuration:\n\nRegistry System\n\nRegistering Middleware\n\nThe MiddlewareRegistry manages all registered middleware:\n\nRegistration Example:\n\nDiscovering Middleware\n\nList all registered middleware:\n\nGet specific middleware:\n\nCheck if middleware is registered:\n\nMiddleware Metadata\n\nEvery middleware must provide metadata:\n\nExample:\n","hierarchy":{"lvl0":"Advanced","lvl1":"Middleware System Architecture","lvl2":"","lvl3":""}},
 {"objectID":"f59113cc43afb8aa74ec51f97fb491b40e5ac27d3bbdaecf1a226437ab0d13a4","title":"Overview","url":"/docs/advanced/middleware-architecture#overview","content":"NeuroLink's middleware system provides a powerful and flexible way to intercept, modify, and enhance AI requests and responses. Middleware enables you to implement cross-cutting concerns like authentication, logging, analytics, content filtering, and auto-evaluation without modifying your core application logic.\n\nWhy Middleware Matters:\nRequest Interception: Modify requests before they reach the AI provider\nResponse Processing: Transform, filter, or validate AI responses\nCross-Cutting Concerns: Implement authentication, logging, rate limiting, and caching in a centralized way\nComposability: Chain multiple middleware components together\nSeparation of Concerns: Keep business logic separate from infrastructure concerns\n\nKey Benefits:\nProduction-ready middleware for common use cases (analytics, guardrails, auto-evaluation)\nFactory pattern for easy middleware management\nPriority-based execution ordering\nProvider-specific conditional execution\nBuilt on NeuroLink's own wrapLanguageModel (src/lib/middleware/wrapLanguageModel.ts), which keeps the familiar wrap-a-model shape without depending on a third-party SDK","hierarchy":{"lvl0":"Advanced","lvl1":"Middleware System Architecture","lvl2":"Overview","lvl3":""}},
 {"objectID":"60b4f50d9fe9e6299393c3a0d190cf5968e65731a535c66e5b17280f20b6bb4f","title":"Request Lifecycle","url":"/docs/advanced/middleware-architecture#request-lifecycle","content":"The middleware system processes requests through four distinct phases:","hierarchy":{"lvl0":"Advanced","lvl1":"Middleware System Architecture","lvl2":"Request Lifecycle","lvl3":""}},
 {"objectID":"d6648ed23071b3b4eaeeb161e0191b57f0aa1d3b718a1ea33c4472f4a822f91a","title":"Phase 1: Pre-Request (transformParams)","url":"/docs/advanced/middleware-architecture#phase-1-pre-request-transformparams","content":"Middleware in this phase runs before the AI provider call, allowing you to:\nValidate input: Check request parameters for validity\nAuthenticate/Authorize: Verify user permissions\nTransform requests: Modify or enrich request parameters\nApply guardrails: Block requests with unsafe content using precall evaluation\nRate limiting: Enforce request quotas\n\nExample Use Cases:\nPrecall guardrails evaluation (blocking unsafe prompts)\nRequest parameter validation\nAdding authentication context\nModifying prompts based on user preferences","hierarchy":{"lvl0":"Advanced","lvl1":"Middleware System Architecture","lvl2":"Phase 1: Pre-Request (transformParams)","lvl3":""}},
@@ -334,7 +335,7 @@
 {"objectID":"20bf1eb219ac6b555aede3c4b1d49ae31ca8ce6f88ab65833ea58f0d563c68e0","title":"Phase 4: Error Handling","url":"/docs/advanced/middleware-architecture#phase-4-error-handling","content":"If an error occurs at any stage, error handling middleware can:\nLog errors: Record error details for debugging\nTransform errors: Convert provider errors to user-friendly messages\nImplement fallbacks: Retry with different providers\nAlert monitoring: Send alerts to monitoring systems\n\nExample Use Cases:\nError logging and tracking\nProvider fallback on failure\nRetry logic with exponential backoff\nUser-friendly error messages","hierarchy":{"lvl0":"Advanced","lvl1":"Middleware System Architecture","lvl2":"Phase 4: Error Handling","lvl3":""}},
 {"objectID":"c1aab1346883c17f6e9da3cf125b3b05a5260dc96ec275774ec50e5e4e0f8671","title":"Execution Order","url":"/docs/advanced/middleware-architecture#execution-order","content":"Middleware executes in priority order, where higher priority values run first:\n\nImportant Notes:\ntransformParams runs before wrapGenerate/wrapStream\nWithin the same priority, registration order determines execution\nMiddleware can be conditionally enabled based on provider, model, or custom logic","hierarchy":{"lvl0":"Advanced","lvl1":"Middleware System Architecture","lvl2":"Execution Order","lvl3":""}},
 {"objectID":"a8f35bb4df1cb02433a72127e19d70d13e65c37f39e82efcf3bcf6e8f87e527e","title":"Chain Configuration","url":"/docs/advanced/middleware-architecture#chain-configuration","content":"Configure which middleware to enable and their order:","hierarchy":{"lvl0":"Advanced","lvl1":"Middleware System Architecture","lvl2":"Chain Configuration","lvl3":""}},
-{"objectID":"4e2d305d305f23f07317c55ef429af7e4f49de1c6b555bf06943a1b82125ecfb","title":"Available Presets","url":"/docs/advanced/middleware-architecture#available-presets","content":"| Preset     | Middleware Enabled     | Use Case               |\n| ---------- | ---------------------- | ---------------------- |\n| default  | Analytics only         | Basic usage tracking   |\n| all      | Analytics + Guardrails | Production with safety |\n| security | Guardrails only        | Security-focused       |\n| Custom     | Your choice            | Define your own        |","hierarchy":{"lvl0":"Advanced","lvl1":"Middleware System Architecture","lvl2":"Available Presets","lvl3":""}},
+{"objectID":"4e2d305d305f23f07317c55ef429af7e4f49de1c6b555bf06943a1b82125ecfb","title":"Available Presets","url":"/docs/advanced/middleware-architecture#available-presets","content":"| Preset     | Middleware Enabled     | Use Case               |\n| ---------- | ---------------------- | ---------------------- |\n| default  | Analytics only         | Basic usage tracking   |\n| all      | Analytics + Guardrails | Production with safety |\n| security | Guardrails only        | Security-focused       |\n| Custom     | Your choice            | Define your own        |\n\nsecurity switches guardrails on and nothing more: it ships no word list and no model filter, so it redacts nothing until badWords, modelFilter or precallEvaluation is configured. A preset, and any other MiddlewareFactoryOptions, can also be set as a default for every call: NEUROLINK_MIDDLEWARE_PRESET (a preset name only), configureGlobalMiddleware() and new NeuroLink({ middleware }), each overridden by the next and finally by the call's own middleware. See Guardrails: Defaults and precedence.","hierarchy":{"lvl0":"Advanced","lvl1":"Middleware System Architecture","lvl2":"Available Presets","lvl3":""}},
 {"objectID":"2ca1866088d786b0302554707dd19943bda8284588e3608359014cd37df2535b","title":"MiddlewareFactory Class","url":"/docs/advanced/middleware-architecture#middlewarefactory-class","content":"The MiddlewareFactory is the central component for managing middleware:","hierarchy":{"lvl0":"Advanced","lvl1":"Middleware System Architecture","lvl2":"MiddlewareFactory Class","lvl3":""}},
 {"objectID":"1a9035a2cc3954f641addd49e21328c92c7328d84ffeea0eba439766fac54125","title":"Creating Middleware Instances","url":"/docs/advanced/middleware-architecture#creating-middleware-instances","content":"Basic Usage:\n\nAdvanced Configuration:","hierarchy":{"lvl0":"Advanced","lvl1":"Middleware System Architecture","lvl2":"Creating Middleware Instances","lvl3":""}},
 {"objectID":"b03a3da8a1b53b12b3c7e81e18e58528c3b4b32a517187311051c1ce954947c7","title":"Registering Middleware","url":"/docs/advanced/middleware-architecture#registering-middleware","content":"The MiddlewareRegistry manages all registered middleware:\n\nRegistration Example:","hierarchy":{"lvl0":"Advanced","lvl1":"Middleware System Architecture","lvl2":"Registering Middleware","lvl3":""}},
@@ -785,7 +786,7 @@
 {"objectID":"6df70184b50cfe765cb953c7f99da86bcc5085d7c1bd9c009ad1d936cd313d54","title":"decide [state]","url":"/docs/cli/commands#decide","content":"Get typed, calibrated judgements from a decision model — the decide inference type, alongside generate and stream. Takes one state plus a map of named typed questions and prints one answer per question; there is no free text anywhere in the response.\n\n`bash","hierarchy":{"lvl0":"Cli","lvl1":"CLI Command Reference","lvl2":"decide [state]","lvl3":""}},
 {"objectID":"0076ed0bb97c3c91b6019b1168d3de94687353bc588fc474812ff9a55bf325d3","title":"State as a positional argument","url":"/docs/cli/commands#state-as-a-positional-argument","content":"npx @juspay/neurolink decide \"Refund request for a damaged item\" \\\n  --questions '{\"urgent\":{\"type\":\"boolean\",\"instructions\":\"Is this urgent?\"}}'","hierarchy":{"lvl0":"Cli","lvl1":"CLI Command Reference","lvl2":"State as a positional argument","lvl3":""}},
 {"objectID":"e7605e932a3b0b4ce109cdb2cbb0d1d5012cb18891ee886ed14d4873544bd69e","title":"State and questions from files, raw JSON output","url":"/docs/cli/commands#state-and-questions-from-files-raw-json-output","content":"npx @juspay/neurolink decide --state-file ticket.json \\\n  --questions-file questions.json --format json","hierarchy":{"lvl0":"Cli","lvl1":"CLI Command Reference","lvl2":"State and questions from files, raw JSON output","lvl3":""}},
-{"objectID":"fa49a31228234b0ba43f669692d1eaa0226061bb36c99c5019d4fc0451af9384","title":"Ask about an image (XOR, Perplexity, Cloudflare Clef and OpenAI Decisions read images); repeat --image for several, or use --video (XOR)","url":"/docs/cli/commands#ask-about-an-image-xor-perplexity-cloudflare-clef-and-openai-decisions-read-images-repeat---image-for-several-or-use---video-xor","content":"npx @juspay/neurolink decide \"What color is this?\" --provider xor \\\n  --image ./photo.png \\\n  --questions '{\"color\":{\"type\":\"choice\",\"instructions\":\"What color is the image?\",\"criteria\":{\"red\":\"red\",\"blue\":\"blue\"}}}'\n\n\n| Option                      | Description                                                                                                                                                                                                                                             |\n| --------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| state                     | The content to judge, as a positional argument (or use --state-file).                                                                                                                                                                                 |\n| --state-file        | Path to a file holding the state (plain text or JSON).                                                                                                                                                                                                  |\n| --questions         | Inline JSON map of questions. Exactly one of this or --questions-file.                                                                                                                                                                                |\n| --questions-file    | Path to a JSON file holding the questions map.","hierarchy":{"lvl0":"Cli","lvl1":"CLI Command Reference","lvl2":"Ask about an image (XOR, Perplexity, Cloudflare Clef and OpenAI Decisions read images); repeat --image for several, or use --video (XOR)","lvl3":""}},
+{"objectID":"fa49a31228234b0ba43f669692d1eaa0226061bb36c99c5019d4fc0451af9384","title":"Ask about an image (XOR, Perplexity, Cloudflare Clef and OpenAI Decisions read images); repeat --image for several, or use --video (XOR)","url":"/docs/cli/commands#ask-about-an-image-xor-perplexity-cloudflare-clef-and-openai-decisions-read-images-repeat---image-for-several-or-use---video-xor","content":"npx @juspay/neurolink decide \"What color is this?\" --provider xor \\\n  --image ./photo.png \\\n  --questions '{\"color\":{\"type\":\"choice\",\"instructions\":\"What color is the image?\",\"criteria\":{\"red\":\"red\",\"blue\":\"blue\"}}}'\n\n\n| Option                      | Description                                                                                                                                                                                                                                                                     |\n| --------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| state                     | The content to judge, as a positional argument (or use --state-file).                                                                                                                                                                                                         |\n| --state-file        | Path to a file holding the state (plain text or JSON).                                                                                                                                                                                                                          |\n| --questions         | Inline JSON map of questions. Exactly one of this or --questions-file.                                                                                                                                                                                                        |\n| --questions-file    | Path to a JSON file holding the questions map.","hierarchy":{"lvl0":"Cli","lvl1":"CLI Command Reference","lvl2":"Ask about an image (XOR, Perplexity, Cloudflare Clef and OpenAI Decisions read images); repeat --image for several, or use --video (XOR)","lvl3":""}},
 {"objectID":"74ddf3bb2432e4cbdfe3a2e78aa95391bd1967ac817b4af735f337554b5b405a","title":"transcribe <file|->","url":"/docs/cli/commands#transcribe","content":"Turn audio into text — the transcribe inference type. Pass a file, or - to read the audio from stdin. With no STT provider configured, the built-in local engine (Whistle) runs on this machine with no key; set NEUROLINK_STT_PROVIDER (and that provider's key) to use another one.\n\n`bash","hierarchy":{"lvl0":"Cli","lvl1":"CLI Command Reference","lvl2":"transcribe <file|->","lvl3":""}},
 {"objectID":"b71aa574620a9889a24a23e9d02ca657151930b6fb1754cb48144df208ef0215","title":"Plain transcription with the default engine","url":"/docs/cli/commands#plain-transcription-with-the-default-engine","content":"npx @juspay/neurolink transcribe meeting.wav","hierarchy":{"lvl0":"Cli","lvl1":"CLI Command Reference","lvl2":"Plain transcription with the default engine","lvl3":""}},
 {"objectID":"688deed48aeb424e78413fcb7603c09c1d4bea2bc3db5c25d4fe8616a23d9a42","title":"Repair names the engine mis-hears, guarded by the decision model, then an LLM rewrite","url":"/docs/cli/commands#repair-names-the-engine-mis-hears-guarded-by-the-decision-model-then-an-llm-rewrite","content":"npx @juspay/neurolink transcribe call.mp3 \\\n  --term \"Acme|akmee,ack me|the company name\" --correct","hierarchy":{"lvl0":"Cli","lvl1":"CLI Command Reference","lvl2":"Repair names the engine mis-hears, guarded by the decision model, then an LLM rewrite","lvl3":""}},
@@ -953,7 +954,7 @@
 {"objectID":"e8f772ab075392c0b4ddf2d907da1d9691928dbdc2c4c0961a603f145aec02f3","title":"Start docs server with stdio transport (default)","url":"/docs/cli/commands#start-docs-server-with-stdio-transport-default","content":"npx @juspay/neurolink docs","hierarchy":{"lvl0":"Cli","lvl1":"CLI Command Reference","lvl2":"Start docs server with stdio transport (default)","lvl3":""}},
 {"objectID":"2688bf05daed20b355836574a11f3879c32211de4b4d79c87d1da17c01d2f361","title":"Start docs server with HTTP transport on custom port","url":"/docs/cli/commands#start-docs-server-with-http-transport-on-custom-port","content":"npx @juspay/neurolink docs --transport http --port 3001\n\n\n| Option        | Alias | Type   | Default | Description                                            |\n| ------------- | ----- | ------ | ------- | ------------------------------------------------------ |\n| --transport | -t  | string | stdio | Transport protocol: stdio (local) or http (remote) |\n| --port      | -p  | number | 3001  | Port for HTTP transport (ignored for stdio)            |\n\n**Note:** Requires the docs site to be built first (cd docs-site && pnpm build`). The server exposes documentation search and retrieval tools via MCP.","hierarchy":{"lvl0":"Cli","lvl1":"CLI Command Reference","lvl2":"Start docs server with HTTP transport on custom port","lvl3":""}},
 {"objectID":"8f49c1467f8e816d39e710ef8143f9ea82f0fb6b7207923b038f16088921ecb5","title":"Troubleshooting","url":"/docs/cli/commands#troubleshooting","content":"| Issue                              | Tip                                                                                                      |\n| ---------------------------------- | -------------------------------------------------------------------------------------------------------- |\n| Unknown argument                 | Check spelling; run command --help for the latest options.                                             |\n| CLI exits immediately              | Upgrade to the newest release or clear old neurolink binaries on PATH.                                 |\n| Provider shows as not-configured | Run neurolink setup --provider <name> or populate .env.                                              |\n| Analytics/evaluation missing       | Ensure both --enableAnalytics/--enableEvaluation and provider credentials for the judge model exist. |\n\nFor advanced workflows (batching, tooling, configuration management) see the relevant guides in the documentation sidebar.","hierarchy":{"lvl0":"Cli","lvl1":"CLI Command Reference","lvl2":"Troubleshooting","lvl3":""}},
-{"objectID":"8119c15fb9af553003230d175f4a2ee502a36966c92c48486de6bae793ea7d1d","title":"Related Features","url":"/docs/cli/commands#related-features","content":"Q4 2025:\nCLI Loop Sessions – Persistent interactive mode with session management\nRedis Conversation Export – Export session history via memory export\nGuardrails Middleware – Content filtering (use --middleware-preset security)\n\nQ3 2025:\nMultimodal Chat – Use --image flag with generate or stream\nAuto Evaluation – Enable with --enableEvaluation\nProvider Orchestration – Automatic fallback and routing\n\nDocumentation:\nSDK API Reference – TypeScript API equivalents\nConfiguration Guide – Environment variables and config files\nTroubleshooting – Detailed error solutions","hierarchy":{"lvl0":"Cli","lvl1":"CLI Command Reference","lvl2":"Related Features","lvl3":""}},
+{"objectID":"8119c15fb9af553003230d175f4a2ee502a36966c92c48486de6bae793ea7d1d","title":"Related Features","url":"/docs/cli/commands#related-features","content":"Q4 2025:\nCLI Loop Sessions – Persistent interactive mode with session management\nRedis Conversation Export – Export session history via memory export\nGuardrails Middleware – Content filtering (the CLI reads NEUROLINK_MIDDLEWARE_PRESET; it has no word-list option)\n\nQ3 2025:\nMultimodal Chat – Use --image flag with generate or stream\nAuto Evaluation – Enable with --enableEvaluation\nProvider Orchestration – Automatic fallback and routing\n\nDocumentation:\nSDK API Reference – TypeScript API equivalents\nConfiguration Guide – Environment variables and config files\nTroubleshooting – Detailed error solutions","hierarchy":{"lvl0":"Cli","lvl1":"CLI Command Reference","lvl2":"Related Features","lvl3":""}},
 {"objectID":"e0f8ded613d2bfce135ac507d9d6d72d0714035e5a4619ae7522926d6929fe66","title":"CLI Examples","url":"/docs/cli/examples","content":"CLI Examples\n\nPractical examples and usage patterns for the NeuroLink CLI.\n\n🚀 Quick Start Examples\n\nBasic Text Generation\n\nProvider Testing\n\n🔧 Development Workflows\n\nCode Generation\n\nDocumentation Generation\n\n📊 Business Use Cases\n\nContent Creation\n\nBusiness Analysis\n\n🔄 Batch Processing\n\nContent Pipeline\n\nCode Review Automation\n\n🎯 Advanced Features\n\nAnalytics and Evaluation\n\nCustom Context\n\n🔍 Debugging and Monitoring\n\nProvider Diagnostics\n\nPerformance Testing\n\n🔧 Integration Examples\n\nShell Scripts\n\nPackage.json Scripts\n\nGitHub Actions\n\n🌐 Production Workflows\n\nContent Management\n\nCode Review Pipeline\n\nMonitoring and Alerts\n\n📈 Performance Optimization\n\nProvider Selection\n\nBatch Optimization\n\n🚨 Error Handling\n\nRobust Scripts\n\nTimeout Handling\n\n📚 Learning and Experimentation\n\nA/B Testing\n\nTemperature Experiments\n\nToken Limit Testing\n\n🔗 Related Resources\nCLI Commands Reference - Complete command documentation\nAdvanced Usage - Power user features\nInstallation Guide - Setup instructions\nEnvironment Variables - Configuration\nTroubleshooting - Common issues","hierarchy":{"lvl0":"Cli","lvl1":"CLI Examples","lvl2":"","lvl3":""}},
 {"objectID":"03625aa61bb31e6c0ba861895a08bad39ca4b9494198f5fceb841ad8bd5cbf84","title":"CLI Examples","url":"/docs/cli/examples#cli-examples","content":"Practical examples and usage patterns for the NeuroLink CLI.","hierarchy":{"lvl0":"Cli","lvl1":"CLI Examples","lvl2":"CLI Examples","lvl3":""}},
 {"objectID":"cbaa79c5e583835d23dcaab7a2e23df19c9e3ec43b92ce0878117717f67f0c5e","title":"Basic Text Generation","url":"/docs/cli/examples#basic-text-generation","content":"`bash","hierarchy":{"lvl0":"Cli","lvl1":"CLI Examples","lvl2":"Basic Text Generation","lvl3":""}},
@@ -2628,8 +2629,8 @@
 {"objectID":"b0a5b78dc4cfcabcd52b58ea94936ee81e97b59b5907f058ccba74881a7aeacd","title":"Context-window filtering","url":"/docs/features/classifier-router-catalog#context-window-filtering","content":"Nothing in routing read maxContextTokens before this. Two independent checks\nnow do:\nrankCatalogue() filters candidates whose contextWindow is below the\n  request's estimated input tokens, applied only when it would leave something\n  behind (an empty result falls back to the unfiltered pool rather than\n  routing nowhere).\nThe classifier's own pick is separately vetoed. Even when jev picks a\n  model directly, fitsRequest() checks whether that specific model's window\n  can hold the request. If it can't, the pick is dropped regardless of\n  confidence — this is a hard provider error, not a degraded answer, because\n  an oversized request against a real model puts that model into ModelPool's\n  permanent (10-year) cooldown. A wrong guess here is not \"less accurate,\" it\n  is unrecoverable for the life of the process.","hierarchy":{"lvl0":"Features","lvl1":"The model catalogue","lvl2":"Context-window filtering","lvl3":""}},
 {"objectID":"e34966644db91aef39b46ef5c2f2fa259e9c828895197c9b2260a07c6e640bce","title":"What this is bad at","url":"/docs/features/classifier-router-catalog#what-this-is-bad-at","content":"Registry quality is coarse, and the catalogue path never shows its own\n  work. Auto-enriched quality is a 3-bucket scale (high/medium/low)\n  before buildModelCatalog() turns it into 1/2/3; two \"high\" models\n  cannot be separated on capability alone. Worse for this page's topic: because\n  the catalogue always populates cost/quality, its rendered line never\n  shows the registry's own price, speed bucket, or \"strong at …\" scores — only\n  the terse capability N / relative cost N form. Declare tiers/quality\n  on a hand-declared pool member for finer control over the number itself; there\n  is no way to get the richer rendering for a catalogue-sourced model.\nA catalogue candidate's \"relative cost\" is real pricing wearing a relative\n  label. buildModelCatalog() seeds it from inputCostPer1K + outputCostPer1K\n  — a small decimal like 0.00074 — not a small integer a host would typically\n  pick for a hand-declared cost. The number is still correct and still never\n  rendered as currency, but it does not compare cleanly against a hand-declared\n  member's cost: 1 in the same pool, since one is a real price and the other\n  is an arbitrary scale.\nUnknown-to-the-registry models rank on relative numbers, not real prices.\n  A model the registry doesn't know (self-hosted, brand-new) keeps whatever\n  cost/quality the host declared, compared only against other declared\n  values on the same relative scale — never rendered as a dollar figure it\n  isn't.\nA wide catalogue is still one choice question, and the cap is hard, not\n  smart, when it does bind. Every model adds tokens to that single question;\n  maxModels is the safeguard, and a model past it would simply never be\n  offered rather than offered with lower priority. Today this is theoretical —\n  the registry is 64 models against a default cap of 120, so nothing is\n  dropped — but the mechanism has no ranking behavior for the day a registry\n  does exceed it.\nPermissive reachability means occasional dead candidates.","hierarchy":{"lvl0":"Features","lvl1":"The model catalogue","lvl2":"What this is bad at","lvl3":""}},
 {"objectID":"2d7354001109ce1051977450cf365d1f2a8e13b345ad206ee23b1f5466228454","title":"See also","url":"/docs/features/classifier-router-catalog#see-also","content":"Classifier Router\nModel routing with a decision model\nThe decide inference type","hierarchy":{"lvl0":"Features","lvl1":"The model catalogue","lvl2":"See also","lvl3":""}},
-{"objectID":"57703ab4e28ce3ad7f4c83d5d3e43f1b1d5541cbb2ac1d6f10f7a057046f0e1b","title":"Model routing with a decision model","url":"/docs/features/classifier-router-jev-strategy","content":"Model routing with a decision model\n\nThe classifier router has a jev strategy: one\ndecision-model round trip answers difficulty,\nrequired capabilities, risk and the model pick simultaneously, with a\ncalibrated confidence on each. This page is the strategy's own mechanics —\nclassifier-router.md covers the router as a whole.\n\nThe degradation contract. With no decision provider configured,\nresolveStrategy() resolves auto to heuristic, exactly as it always did.\nclassifyJev() itself never throws: any failure, timeout, or malformed answer\nfalls back to classifyHeuristic(). Configuring a decision provider (for example TYPESAFE_API_KEY or\nAI_GATEWAY_API_KEY) upgrades routing; it cannot make routing worse than before\nthe key existed. The same holds for PERPLEXITY_API_KEY, which is also the\nPerplexity text provider's key, and for CLOUDFLARE_API_KEY with\nCLOUDFLARE_ACCOUNT_ID, which are also the Cloudflare Workers AI text provider's\ntoken and account id, and for SUBQ_API_KEY, and for OPENAI_API_KEY, which is also the OpenAI chat\nprovider's key (for a host with only that key auto now resolves to jev,\nanswered by OpenAI Decisions as the last-resort decision provider, unless OPENAI_BASE_URL,\nOPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com, which makes it a gateway's key). \"Configured\" counts the environment, the instance\ncredentials, and the request's own per-call credentials — a host that holds\nno decision key but passes a tenant's credentials: { laya } on a generate()\ncall routes that call with jev, and the decision goes to the tenant's account.\n\nWhat goes into the state\n\nclassifyJev() sends the request truncated to 8000 characters, plus four\nsignals the caller already has on hand:\n\nAlongside it, one batch asks: difficulty (a choice over the five tiers),\nneeds_vision / needs_tools / needs_reasoning (boolean), risky\n(boolean), context (a score — see\nper-request context budget), and, only when the\npool has more than one member, model (a choice over the pool, rendered by\nthe catalogue). All of this rides in\none request. On TypeSafe that takes ~400ms, because latency is flat in question\ncount; on Perplexity each further question adds about 65 ms, up to 128 (see\nits guide); on\nCloudflare Clef a small request took 0.3 to 1.0 s on 2026-10-03 (see\nits guide). See\nthe batching rule.\n\nNeuroLink refuses a state it estimates at more than 16,000 tokens (about 60,000\ncharacters of prose) with max_tokens_exceeded, to keep Clef inside what it was\nseen to read in full and inside a timeout a fail-open consumer can afford. A\nlonger request is refused there, and the heuristic classifier's tier stands, as it\ndoes on any failure.\n\nThe difficulty rubric\n\nFive tiers, ordered easiest to hardest, worded about the shape of the work —\nthe model is never told a provider or model name:\n\n| Tier       | Criterion                                                                                                                                             |\n| ---------- | ----------------------------------------------------------------------------------------------------------------------------------------------------- |\n| trivial  | Mechanical and local: rename a symbol, fix a typo, add an import, run one named command, or answer something already stated.                          |\n| simple   | A small localised change or a direct factual answer. One file, one obvious approach.                                                                  |\n| moderate | Ordinary engineering: implement a well-specified change across a few files, write tests, fix a clearly described bug, review a small diff.            |\n| hard     | Deep reasoning: architecture and design, debugging a failure whose cause is unknown, security analysis, concurrency, cross-system refactors.          |\n| expert   | Frontier-level work: ambiguous requirements, novel design with no established pattern, or analysis where a wrong answer is costly and hard to detect. |\n\nAsymmetric confidence bars, and why they differ\n\nA verdict harder than the neutral tier (moderate) spends more if wrong; a\nverdict easier than neutral spends less if wrong. Those are not the same\nmistake:\nRouting a simple task to an expensive model wastes money. Cheap to be wrong\n  about, so the bar to route up is low: minUpgradeConfidence defaults to\n  0.3.\nRouting a hard task to a weak model produces a wrong answer. Expensive to be\n  wrong about, so the bar to route down is high: minDowngradeConfidence\n  defaults to 0.6.\n\nBelow the applicable bar, the difficulty verdict is discarded and the heuristic\nclassifier's tier stands instead — not a downgraded jev answer, the ordinary\nzero-cost fallback.\n\nOne case skips both bars: a risky reading above 0.7 forces the tier to at\nleast hard, unconditionally. Risk can only ever raise the tier, never lower\none, and it is not itself gated by confidence.\n\nAsk about the act, not the subject\n\nThe risk question is not \"does this touch production, mon","hierarchy":{"lvl0":"Features","lvl1":"Model routing with a decision model","lvl2":"","lvl3":""}},
-{"objectID":"bd360e5abf4e93403d1fb23a27c946fed1b86d8d8863a98519338332ca726639","title":"Model routing with a decision model","url":"/docs/features/classifier-router-jev-strategy#model-routing-with-a-decision-model","content":"The classifier router has a jev strategy: one\ndecision-model round trip answers difficulty,\nrequired capabilities, risk and the model pick simultaneously, with a\ncalibrated confidence on each. This page is the strategy's own mechanics —\nclassifier-router.md covers the router as a whole.\n\nThe degradation contract. With no decision provider configured,\nresolveStrategy() resolves auto to heuristic, exactly as it always did.\nclassifyJev() itself never throws: any failure, timeout, or malformed answer\nfalls back to classifyHeuristic(). Configuring a decision provider (for example TYPESAFE_API_KEY or\nAI_GATEWAY_API_KEY) upgrades routing; it cannot make routing worse than before\nthe key existed. The same holds for PERPLEXITY_API_KEY, which is also the\nPerplexity text provider's key, and for CLOUDFLARE_API_KEY with\nCLOUDFLARE_ACCOUNT_ID, which are also the Cloudflare Workers AI text provider's\ntoken and account id, and for SUBQ_API_KEY, and for OPENAI_API_KEY, which is also the OpenAI chat\nprovider's key (for a host with only that key auto now resolves to jev,\nanswered by OpenAI Decisions as the last-resort decision provider, unless OPENAI_BASE_URL,\nOPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com, which makes it a gateway's key). \"Configured\" counts the environment, the instance\ncredentials, and the request's own per-call credentials — a host that holds\nno decision key but passes a tenant's credentials: { laya } on a generate()\ncall routes that call with jev, and the decision goes to the tenant's account.","hierarchy":{"lvl0":"Features","lvl1":"Model routing with a decision model","lvl2":"Model routing with a decision model","lvl3":""}},
+{"objectID":"57703ab4e28ce3ad7f4c83d5d3e43f1b1d5541cbb2ac1d6f10f7a057046f0e1b","title":"Model routing with a decision model","url":"/docs/features/classifier-router-jev-strategy","content":"Model routing with a decision model\n\nThe classifier router has a jev strategy: one\ndecision-model round trip answers difficulty,\nrequired capabilities, risk and the model pick simultaneously, with a\ncalibrated confidence on each. This page is the strategy's own mechanics —\nclassifier-router.md covers the router as a whole.\n\nThe degradation contract. With no decision provider configured,\nresolveStrategy() resolves auto to heuristic, exactly as it always did.\nclassifyJev() itself never throws: any failure, timeout, or malformed answer\nfalls back to classifyHeuristic(). Configuring a decision provider (for example TYPESAFE_API_KEY or\nAI_GATEWAY_API_KEY) upgrades routing; it cannot make routing worse than before\nthe key existed. The same holds for PERPLEXITY_API_KEY, which is also the\nPerplexity text provider's key, and for CLOUDFLARE_API_KEY with\nCLOUDFLARE_ACCOUNT_ID, which are also the Cloudflare Workers AI text provider's\ntoken and account id, and for SUBQ_API_KEY, and for OPENAI_API_KEY, which is also the OpenAI chat\nprovider's key (for a host with only that key auto now resolves to jev,\nanswered by OpenAI Decisions, the default decision provider when none ahead of it is configured, unless OPENAI_BASE_URL,\nOPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com, which makes it a gateway's key), and for SCALEDOWN_API_KEY, which is ScaleDown's own and counts by itself (with only that key auto resolves to jev, answered by ScaleDown). \"Configured\" counts the environment, the instance\ncredentials, and the request's own per-call credentials — a host that holds\nno decision key but passes a tenant's credentials: { laya } on a generate()\ncall routes that call with jev, and the decision goes to the tenant's account.\n\nWhat goes into the state\n\nclassifyJev() sends the request truncated to 8000 characters, plus four\nsignals the caller already has on hand:\n\nAlongside it, one batch asks: difficulty (a choice over the five tiers),\nneeds_vision / needs_tools / needs_reasoning (boolean), risky\n(boolean), context (a score — see\nper-request context budget), and, only when the\npool has more than one member, model (a choice over the pool, rendered by\nthe catalogue). All of this rides in\none request. On TypeSafe that takes ~400ms, because latency is flat in question\ncount; on Perplexity each further question adds about 65 ms, up to 128 (see\nits guide); on\nCloudflare Clef a small request took 0.3 to 1.0 s on 2026-10-03 (see\nits guide). See\nthe batching rule.\n\nNeuroLink refuses a state it estimates at more than 16,000 tokens (about 60,000\ncharacters of prose) with max_tokens_exceeded, to keep Clef inside what it was\nseen to read in full and inside a timeout a fail-open consumer can afford. A\nlonger request is refused there, and the heuristic classifier's tier stands, as it\ndoes on any failure.\n\nThe difficulty rubric\n\nFive tiers, ordered easiest to hardest, worded about the shape of the work —\nthe model is never told a provider or model name:\n\n| Tier       | Criterion                                                                                                                                             |\n| ---------- | ----------------------------------------------------------------------------------------------------------------------------------------------------- |\n| trivial  | Mechanical and local: rename a symbol, fix a typo, add an import, run one named command, or answer something already stated.                          |\n| simple   | A small localised change or a direct factual answer. One file, one obvious approach.                                                                  |\n| moderate | Ordinary engineering: implement a well-specified change across a few files, write tests, fix a clearly described bug, review a small diff.            |\n| hard     | Deep reasoning: architecture and design, debugging a failure whose cause is unknown, security analysis, concurrency, cross-system refactors.          |\n| expert   | Frontier-level work: ambiguous requirements, novel design with no established pattern, or analysis where a wrong answer is costly and hard to detect. |\n\nAsymmetric confidence bars, and why they differ\n\nA verdict harder than the neutral tier (moderate) spends more if wrong; a\nverdict easier than neutral spends less if wrong. Those are not the same\nmistake:\nRouting a simple task to an expensive model wastes money. Cheap to be wrong\n  about, so the bar to route up is low: minUpgradeConfidence defaults to\n  0.3.\nRouting a hard task to a weak model produces a wrong answer. Expensive to be\n  wrong about, so the bar to route down is high: minDowngradeConfidence\n  defaults to 0.6.\n\nBelow the applicable bar, the difficulty verdict is discarded and the heuristic\nclassifier's tier stands instead — not a downgraded jev answer, the ordinary\nzero-cost fallback.\n\nOne case skips both bars: a risky reading above 0.7 forces the tier to at\nleast hard, unconditionally. Risk can only ever ","hierarchy":{"lvl0":"Features","lvl1":"Model routing with a decision model","lvl2":"","lvl3":""}},
+{"objectID":"bd360e5abf4e93403d1fb23a27c946fed1b86d8d8863a98519338332ca726639","title":"Model routing with a decision model","url":"/docs/features/classifier-router-jev-strategy#model-routing-with-a-decision-model","content":"The classifier router has a jev strategy: one\ndecision-model round trip answers difficulty,\nrequired capabilities, risk and the model pick simultaneously, with a\ncalibrated confidence on each. This page is the strategy's own mechanics —\nclassifier-router.md covers the router as a whole.\n\nThe degradation contract. With no decision provider configured,\nresolveStrategy() resolves auto to heuristic, exactly as it always did.\nclassifyJev() itself never throws: any failure, timeout, or malformed answer\nfalls back to classifyHeuristic(). Configuring a decision provider (for example TYPESAFE_API_KEY or\nAI_GATEWAY_API_KEY) upgrades routing; it cannot make routing worse than before\nthe key existed. The same holds for PERPLEXITY_API_KEY, which is also the\nPerplexity text provider's key, and for CLOUDFLARE_API_KEY with\nCLOUDFLARE_ACCOUNT_ID, which are also the Cloudflare Workers AI text provider's\ntoken and account id, and for SUBQ_API_KEY, and for OPENAI_API_KEY, which is also the OpenAI chat\nprovider's key (for a host with only that key auto now resolves to jev,\nanswered by OpenAI Decisions, the default decision provider when none ahead of it is configured, unless OPENAI_BASE_URL,\nOPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com, which makes it a gateway's key), and for SCALEDOWN_API_KEY, which is ScaleDown's own and counts by itself (with only that key auto resolves to jev, answered by ScaleDown). \"Configured\" counts the environment, the instance\ncredentials, and the request's own per-call credentials — a host that holds\nno decision key but passes a tenant's credentials: { laya } on a generate()\ncall routes that call with jev, and the decision goes to the tenant's account.","hierarchy":{"lvl0":"Features","lvl1":"Model routing with a decision model","lvl2":"Model routing with a decision model","lvl3":""}},
 {"objectID":"a6b27c365174a477eafb6835b9ab5990c7434acd58cb21534f8419ea21c63dd5","title":"What goes into the state","url":"/docs/features/classifier-router-jev-strategy#what-goes-into-the-state","content":"classifyJev() sends the request truncated to 8000 characters, plus four\nsignals the caller already has on hand:\n\nAlongside it, one batch asks: difficulty (a choice over the five tiers),\nneeds_vision / needs_tools / needs_reasoning (boolean), risky\n(boolean), context (a score — see\nper-request context budget), and, only when the\npool has more than one member, model (a choice over the pool, rendered by\nthe catalogue). All of this rides in\none request. On TypeSafe that takes ~400ms, because latency is flat in question\ncount; on Perplexity each further question adds about 65 ms, up to 128 (see\nits guide); on\nCloudflare Clef a small request took 0.3 to 1.0 s on 2026-10-03 (see\nits guide). See\nthe batching rule.\n\nNeuroLink refuses a state it estimates at more than 16,000 tokens (about 60,000\ncharacters of prose) with max_tokens_exceeded, to keep Clef inside what it was\nseen to read in full and inside a timeout a fail-open consumer can afford. A\nlonger request is refused there, and the heuristic classifier's tier stands, as it\ndoes on any failure.","hierarchy":{"lvl0":"Features","lvl1":"Model routing with a decision model","lvl2":"What goes into the state","lvl3":""}},
 {"objectID":"ed3d90e1e698e522c2fb4cbd27b81d89a769c3d860bbbb85d004ef5cf1d1090d","title":"The difficulty rubric","url":"/docs/features/classifier-router-jev-strategy#the-difficulty-rubric","content":"Five tiers, ordered easiest to hardest, worded about the shape of the work —\nthe model is never told a provider or model name:\n\n| Tier       | Criterion                                                                                                                                             |\n| ---------- | ----------------------------------------------------------------------------------------------------------------------------------------------------- |\n| trivial  | Mechanical and local: rename a symbol, fix a typo, add an import, run one named command, or answer something already stated.                          |\n| simple   | A small localised change or a direct factual answer. One file, one obvious approach.                                                                  |\n| moderate | Ordinary engineering: implement a well-specified change across a few files, write tests, fix a clearly described bug, review a small diff.            |\n| hard     | Deep reasoning: architecture and design, debugging a failure whose cause is unknown, security analysis, concurrency, cross-system refactors.          |\n| expert   | Frontier-level work: ambiguous requirements, novel design with no established pattern, or analysis where a wrong answer is costly and hard to detect. |","hierarchy":{"lvl0":"Features","lvl1":"Model routing with a decision model","lvl2":"The difficulty rubric","lvl3":""}},
 {"objectID":"2bcf89adce5486bff7cc01f6916f9bf26080de8996ec2277df6e15eb3eb264e3","title":"Asymmetric confidence bars, and why they differ","url":"/docs/features/classifier-router-jev-strategy#asymmetric-confidence-bars-and-why-they-differ","content":"A verdict harder than the neutral tier (moderate) spends more if wrong; a\nverdict easier than neutral spends less if wrong. Those are not the same\nmistake:\nRouting a simple task to an expensive model wastes money. Cheap to be wrong\n  about, so the bar to route up is low: minUpgradeConfidence defaults to\n  0.3.\nRouting a hard task to a weak model produces a wrong answer. Expensive to be\n  wrong about, so the bar to route down is high: minDowngradeConfidence\n  defaults to 0.6.\n\nBelow the applicable bar, the difficulty verdict is discarded and the heuristic\nclassifier's tier stands instead — not a downgraded jev answer, the ordinary\nzero-cost fallback.\n\nOne case skips both bars: a risky reading above 0.7 forces the tier to at\nleast hard, unconditionally. Risk can only ever raise the tier, never lower\none, and it is not itself gated by confidence.","hierarchy":{"lvl0":"Features","lvl1":"Model routing with a decision model","lvl2":"Asymmetric confidence bars, and why they differ","lvl3":""}},
@@ -2637,10 +2638,10 @@
 {"objectID":"d90c4af28403b7660202646864195bb2ff353a7f7a29ae72f01fd8aa997a39dd","title":"The model pick faces a bar too — but a different one","url":"/docs/features/classifier-router-jev-strategy#the-model-pick-faces-a-bar-too-but-a-different-one","content":"When the pool has more than one member, jev is also asked to choose directly:\n_\"Which of these models is the cheapest one that can still complete this\nrequest correctly?\"_ — over\nthe rendered candidate lines.\nThat pick is reported with its own confidence, separate from the difficulty\nconfidence, and the router decides which bar applies:\n\nPicking something costlier than what the difficulty tier would have picked\non its own risks only spending more than necessary, so it clears the low\nupgrade bar. Picking something cheaper risks handing the task to a model\nthat cannot do it, so it must clear the high downgrade bar. Agreeing with the\ntier needs no bar at all. If the pick fails its bar, it is dropped — logged as\n\"classifier pick dropped — below its bar\" — and the tier's own ranked list is\nused instead.\n\nThis is a genuinely different question from the difficulty asymmetry above:\nthat one asks whether the tier is trustworthy; this one asks whether the\nspecific model choice, once a tier is settled, is trustworthy — and the two\ncan point in opposite directions (a confident-enough \"hard\" verdict whose model\npick still misses its own, stricter bar).\n\nA pick that cannot physically hold the request is not gated at all — it is\ndropped outright, regardless of confidence, because that is a hard provider\nerror rather than a degraded answer: see\nthe catalogue's context-window filtering.\n\nThe llm strategy's picks are exempt from both bars: it reports no confidence\nfor its pick, so a bar would either always pass or always fail. Its picks are\nhonoured exactly as they were before jev existed — imposing a bar here would\nsilently change an unrelated, already-shipped strategy.","hierarchy":{"lvl0":"Features","lvl1":"Model routing with a decision model","lvl2":"The model pick faces a bar too — but a different one","lvl3":""}},
 {"objectID":"4553a84789531ca2e1bad075175d684e27e4ab3fcbb009062048cdf11765c2fb","title":"What this is bad at","url":"/docs/features/classifier-router-jev-strategy#what-this-is-bad-at","content":"It cannot tell you why. The reason field is a debug string built from\n  the numbers, not an explanation the model gave. If you need an auditable\n  rationale for a routing decision, this is the wrong tool.\nA close call still routes somewhere. A 0.29 upgrade verdict and a 0.61\n  downgrade verdict are both one hundredth of a point from the bar, and both\n  fall all the way back to the heuristic tier rather than to \"the second most\n  likely tier.\" There is no partial credit.\nThe risk question is a single boolean. It cannot express \"risky, but\n  only mildly\" — anything past 0.7 jumps straight to hard, whatever the\n  actual severity.\nIt shares the base model's general limits. Literal reading, no\n  arithmetic, and state relevance affecting accuracy all apply here exactly as\n  described in what decide is bad at.\nThe pool still has to exist. jev chooses among what you declared (or\n  what the catalogue built); it\n  cannot invent a model you have no credentials for.","hierarchy":{"lvl0":"Features","lvl1":"Model routing with a decision model","lvl2":"What this is bad at","lvl3":""}},
 {"objectID":"ab4d25365369a722077741378e60be6e3ef36f612ed7a85340c93cf9d47223da","title":"See also","url":"/docs/features/classifier-router-jev-strategy#see-also","content":"The decide inference type\nClassifier Router\nThe model catalogue\nPer-request context budget","hierarchy":{"lvl0":"Features","lvl1":"Model routing with a decision model","lvl2":"See also","lvl3":""}},
-{"objectID":"3f990fb1489f99d505b19941e00fd05b9368a5a91745c83666639aac85865343","title":"Classifier Router","url":"/docs/features/classifier-router","content":"Classifier Router\n\nStatus: Stable | Availability: SDK + CLI | Opt-in (disabled by default)\n\nOverview\n\nThe Classifier Router lets NeuroLink decide, per request, which model to use (and, optionally, which tools to expose) from a pool you declare — routing hard/complex tasks to more capable models and easy tasks to cheaper, faster ones. You give it a pool of models; it classifies the incoming prompt and switches to the best one transparently before the call runs.\n\nIt is opt-in (classifierRouter.enabled defaults to false), fails open (any classifier/selection error leaves the call exactly as it would have been), and is fully backward compatible — a default new NeuroLink() is unchanged.\n\nTypical use cases:\nCost optimization — send \"hi\" to a cheap model and a multi-step architecture question to a powerful one, automatically.\nLatency optimization — keep simple turns on fast models.\nCustom / self-hosted fleets — route across LiteLLM, OpenAI-compatible, or Ollama models that aren't in any registry.\nPer-difficulty tool scoping — expose fewer tools for trivial tasks.\n\nHow it works\n\nEach request flows through two stages:\nClassify — produce a difficulty bucket (trivial | simple | moderate | hard | expert) plus optional requiredCapabilities and tool hints. Four strategies:\nauto (default): resolves to jev when a decision provider is configured (TYPESAFE_API_KEY or AI_GATEWAY_API_KEY, LAYA_API_KEY with LAYA_BASE_URL, XOR_API_KEY with XOR_BASE_URL, SUBQ_API_KEY, PERPLEXITY_API_KEY, CLOUDFLARE_API_KEY with CLOUDFLARE_ACCOUNT_ID, or OPENAI_API_KEY), and heuristic otherwise. Setting a key therefore upgrades routing with no code change; without one, behaviour is exactly as it was. PERPLEXITY_API_KEY is also the Perplexity text provider's key, so it counts here too, and CLOUDFLARE_API_KEY with CLOUDFLARE_ACCOUNT_ID are also the Cloudflare Workers AI text provider's token and account id, so they count here as well; OPENAI_API_KEY is also the OpenAI chat provider's key, so for a host with only that key auto now resolves to jev, answered by OpenAI Decisions as the last-resort decision provider (a paid call, though routing still falls back to the heuristic on any failure), unless OPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com: that key is taken to be a gateway's and auto stays on heuristic. The gate counts the credentials given to the NeuroLink constructor, the environment and the request's own per-call credentials, so a credentials.subq.apiKey (or credentials.laya, credentials.xor, credentials.perplexityDecider, credentials.cloudflareClef, credentials.openaiDecisions and the other decision providers' slices) passed on a single call can make auto pick jev rather than heuristic, and the decision then goes to that account.\nheuristic: zero-cost keyword/length scoring of the prompt text. No LLM call, fully deterministic, provider-agnostic.\nllm: a cheap \"classifier model\" reads the prompt and returns a difficulty — and, when given your pool, picks a model directly by id.\njev: a decision model (the first configured decision provider, in the order TypeSafe's Jev, Laya, XOR, SubQ, Perplexity, Cloudflare Clef, OpenAI Decisions) answers difficulty, required capabilities and the model pick in one request, with a confidence for each answer. On TypeSafe's Jev that request takes ~400 ms and the confidence is calibrated; the Perplexity guide gives its latency and the confidence it reports, and the Cloudflare Clef guide gives its latency and its state limit (16,000 estimated tokens, so a longer prompt is refused and the heuristic stands). Verdicts that miss the applicable confidence bar (minUpgradeConfidence 0.3 to route up, minDowngradeConfidence 0.6 to route down) fall through to the heuristic rather than acting on a guess. See the decide inference type.\nSelect — turn that into a concrete { provider, model, region } from your pool, optionally narrowing tools.\n\nThe router runs before the provider/model is constructed (it reuses the same pre-call seam as requestRouter). It is skipped when the caller pinned both provider and model, or when a modelPool is configured (the pool owns selection).\n\nQuick start (heuristic, SDK)\n\nDefining \"which model for which case\"\n\nThe router resolves a model using the first of these that applies:\n\n| #   | Mechanism              | How you define it                                                                  | Best for                              |\n| --- | ---------------------- | ---------------------------------------------------------------------------------- | ------------------------------------- |\n| 1   | LLM direct pick    | classifier: \"llm\" + a description on each pool member                          | Custom/registry-less models; smartest |\n| 2   | tierMap          | Explicit difficulty → members map                                                | Full, deterministic control           |\n| 3   | Per-member tiers | tiers: [\"hard\",\"expert\"] on a ","hierarchy":{"lvl0":"Features","lvl1":"Classifier Router","lvl2":"","lvl3":""}},
+{"objectID":"3f990fb1489f99d505b19941e00fd05b9368a5a91745c83666639aac85865343","title":"Classifier Router","url":"/docs/features/classifier-router","content":"Classifier Router\n\nStatus: Stable | Availability: SDK + CLI | Opt-in (disabled by default)\n\nOverview\n\nThe Classifier Router lets NeuroLink decide, per request, which model to use (and, optionally, which tools to expose) from a pool you declare — routing hard/complex tasks to more capable models and easy tasks to cheaper, faster ones. You give it a pool of models; it classifies the incoming prompt and switches to the best one transparently before the call runs.\n\nIt is opt-in (classifierRouter.enabled defaults to false), fails open (any classifier/selection error leaves the call exactly as it would have been), and is fully backward compatible — a default new NeuroLink() is unchanged.\n\nTypical use cases:\nCost optimization — send \"hi\" to a cheap model and a multi-step architecture question to a powerful one, automatically.\nLatency optimization — keep simple turns on fast models.\nCustom / self-hosted fleets — route across LiteLLM, OpenAI-compatible, or Ollama models that aren't in any registry.\nPer-difficulty tool scoping — expose fewer tools for trivial tasks.\n\nHow it works\n\nEach request flows through two stages:\nClassify — produce a difficulty bucket (trivial | simple | moderate | hard | expert) plus optional requiredCapabilities and tool hints. Four strategies:\nauto (default): resolves to jev when a decision provider is configured (TYPESAFE_API_KEY or AI_GATEWAY_API_KEY, LAYA_API_KEY with LAYA_BASE_URL, XOR_API_KEY with XOR_BASE_URL, SUBQ_API_KEY, PERPLEXITY_API_KEY, CLOUDFLARE_API_KEY with CLOUDFLARE_ACCOUNT_ID, OPENAI_API_KEY, or SCALEDOWN_API_KEY), and heuristic otherwise. Setting a key therefore upgrades routing with no code change; without one, behaviour is exactly as it was. PERPLEXITY_API_KEY is also the Perplexity text provider's key, so it counts here too, and CLOUDFLARE_API_KEY with CLOUDFLARE_ACCOUNT_ID are also the Cloudflare Workers AI text provider's token and account id, so they count here as well; OPENAI_API_KEY is also the OpenAI chat provider's key, so for a host with only that key auto now resolves to jev, answered by OpenAI Decisions, the default decision provider when none ahead of it is configured (a paid call, though routing still falls back to the heuristic on any failure), unless OPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com: that key is taken to be a gateway's and auto stays on heuristic. SCALEDOWN_API_KEY is ScaleDown's own key and counts by itself: with only that key auto resolves to jev, answered by ScaleDown (also a paid call, with the same fallback). The gate counts the credentials given to the NeuroLink constructor, the environment and the request's own per-call credentials, so a credentials.subq.apiKey (or credentials.laya, credentials.xor, credentials.perplexityDecider, credentials.cloudflareClef, credentials.openaiDecisions, credentials.scaledown and the other decision providers' slices) passed on a single call can make auto pick jev rather than heuristic, and the decision then goes to that account.\nheuristic: zero-cost keyword/length scoring of the prompt text. No LLM call, fully deterministic, provider-agnostic.\nllm: a cheap \"classifier model\" reads the prompt and returns a difficulty — and, when given your pool, picks a model directly by id.\njev: a decision model (the first configured decision provider, in the order TypeSafe's Jev, Laya, XOR, SubQ, Perplexity, Cloudflare Clef, OpenAI Decisions, ScaleDown) answers difficulty, required capabilities and the model pick in one request, with a confidence for each answer. On TypeSafe's Jev that request takes ~400 ms and the confidence is calibrated; the Perplexity guide gives its latency and the confidence it reports, and the Cloudflare Clef guide gives its latency and its state limit (16,000 estimated tokens, so a longer prompt is refused and the heuristic stands). Verdicts that miss the applicable confidence bar (minUpgradeConfidence 0.3 to route up, minDowngradeConfidence 0.6 to route down) fall through to the heuristic rather than acting on a guess. See the decide inference type.\nSelect — turn that into a concrete { provider, model, region } from your pool, optionally narrowing tools.\n\nThe router runs before the provider/model is constructed (it reuses the same pre-call seam as requestRouter). It is skipped when the caller pinned both provider and model, or when a modelPool is configured (the pool owns selection).\n\nQuick start (heuristic, SDK)\n\nDefining \"which model for which case\"\n\nThe router resolves a model using the first of these that applies:\n\n| #   | Mechanism              | How you define it                                                                  | Best for                              |\n| --- | ---------------------- | ---------------------------------------------------------------------------------- | ------------------------------------- |\n| 1   | LLM direct pick    | classifier: \"llm\" + a description on each pool member                      ","hierarchy":{"lvl0":"Features","lvl1":"Classifier Router","lvl2":"","lvl3":""}},
 {"objectID":"b8bb35b0db438959a0a557c68b22afa16480504b96128260a8947ba3c8a81f9f","title":"Classifier Router","url":"/docs/features/classifier-router#classifier-router","content":"Status: Stable | Availability: SDK + CLI | Opt-in (disabled by default)","hierarchy":{"lvl0":"Features","lvl1":"Classifier Router","lvl2":"Classifier Router","lvl3":""}},
 {"objectID":"543ddabeefd7754f6047758090a6a0ac6f517c98b9afb146645b9073ad75c5ee","title":"Overview","url":"/docs/features/classifier-router#overview","content":"The Classifier Router lets NeuroLink decide, per request, which model to use (and, optionally, which tools to expose) from a pool you declare — routing hard/complex tasks to more capable models and easy tasks to cheaper, faster ones. You give it a pool of models; it classifies the incoming prompt and switches to the best one transparently before the call runs.\n\nIt is opt-in (classifierRouter.enabled defaults to false), fails open (any classifier/selection error leaves the call exactly as it would have been), and is fully backward compatible — a default new NeuroLink() is unchanged.\n\nTypical use cases:\nCost optimization — send \"hi\" to a cheap model and a multi-step architecture question to a powerful one, automatically.\nLatency optimization — keep simple turns on fast models.\nCustom / self-hosted fleets — route across LiteLLM, OpenAI-compatible, or Ollama models that aren't in any registry.\nPer-difficulty tool scoping — expose fewer tools for trivial tasks.","hierarchy":{"lvl0":"Features","lvl1":"Classifier Router","lvl2":"Overview","lvl3":""}},
-{"objectID":"abdbdbfcfd51c6b40ef9cdf848fa23aeb87e4e4513c71572f89ee754be41d834","title":"How it works","url":"/docs/features/classifier-router#how-it-works","content":"Each request flows through two stages:\nClassify — produce a difficulty bucket (trivial | simple | moderate | hard | expert) plus optional requiredCapabilities and tool hints. Four strategies:\nauto (default): resolves to jev when a decision provider is configured (TYPESAFE_API_KEY or AI_GATEWAY_API_KEY, LAYA_API_KEY with LAYA_BASE_URL, XOR_API_KEY with XOR_BASE_URL, SUBQ_API_KEY, PERPLEXITY_API_KEY, CLOUDFLARE_API_KEY with CLOUDFLARE_ACCOUNT_ID, or OPENAI_API_KEY), and heuristic otherwise. Setting a key therefore upgrades routing with no code change; without one, behaviour is exactly as it was. PERPLEXITY_API_KEY is also the Perplexity text provider's key, so it counts here too, and CLOUDFLARE_API_KEY with CLOUDFLARE_ACCOUNT_ID are also the Cloudflare Workers AI text provider's token and account id, so they count here as well; OPENAI_API_KEY is also the OpenAI chat provider's key, so for a host with only that key auto now resolves to jev, answered by OpenAI Decisions as the last-resort decision provider (a paid call, though routing still falls back to the heuristic on any failure), unless OPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com: that key is taken to be a gateway's and auto stays on heuristic. The gate counts the credentials given to the NeuroLink constructor, the environment and the request's own per-call credentials, so a credentials.subq.apiKey (or credentials.laya, credentials.xor, credentials.perplexityDecider, credentials.cloudflareClef, credentials.openaiDecisions and the other decision providers' slices) passed on a single call can make auto pick jev rather than heuristic, and the decision then goes to that account.\nheuristic: zero-cost keyword/length scoring of the prompt text. No LLM call, fully deterministic, provider-agnostic.\nllm: a cheap \"classifier model\" reads the prompt and returns a difficulty — and, when given your pool, picks a model directly by id.","hierarchy":{"lvl0":"Features","lvl1":"Classifier Router","lvl2":"How it works","lvl3":""}},
+{"objectID":"abdbdbfcfd51c6b40ef9cdf848fa23aeb87e4e4513c71572f89ee754be41d834","title":"How it works","url":"/docs/features/classifier-router#how-it-works","content":"Each request flows through two stages:\nClassify — produce a difficulty bucket (trivial | simple | moderate | hard | expert) plus optional requiredCapabilities and tool hints. Four strategies:\nauto (default): resolves to jev when a decision provider is configured (TYPESAFE_API_KEY or AI_GATEWAY_API_KEY, LAYA_API_KEY with LAYA_BASE_URL, XOR_API_KEY with XOR_BASE_URL, SUBQ_API_KEY, PERPLEXITY_API_KEY, CLOUDFLARE_API_KEY with CLOUDFLARE_ACCOUNT_ID, OPENAI_API_KEY, or SCALEDOWN_API_KEY), and heuristic otherwise. Setting a key therefore upgrades routing with no code change; without one, behaviour is exactly as it was. PERPLEXITY_API_KEY is also the Perplexity text provider's key, so it counts here too, and CLOUDFLARE_API_KEY with CLOUDFLARE_ACCOUNT_ID are also the Cloudflare Workers AI text provider's token and account id, so they count here as well; OPENAI_API_KEY is also the OpenAI chat provider's key, so for a host with only that key auto now resolves to jev, answered by OpenAI Decisions, the default decision provider when none ahead of it is configured (a paid call, though routing still falls back to the heuristic on any failure), unless OPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com: that key is taken to be a gateway's and auto stays on heuristic. SCALEDOWN_API_KEY is ScaleDown's own key and counts by itself: with only that key auto resolves to jev, answered by ScaleDown (also a paid call, with the same fallback). The gate counts the credentials given to the NeuroLink constructor, the environment and the request's own per-call credentials, so a credentials.subq.apiKey (or credentials.laya, credentials.xor, credentials.perplexityDecider, credentials.cloudflareClef, credentials.openaiDecisions, credentials.scaledown and the other decision providers' slices) passed on a single call can make auto pick jev rather than heuristic, and the decision then goes to that account.","hierarchy":{"lvl0":"Features","lvl1":"Classifier Router","lvl2":"How it works","lvl3":""}},
 {"objectID":"08f69ffa1f39bf410a5c236efa5f21f51efc93034747ddab6d761f4b1615d7df","title":"Defining \"which model for which case\"","url":"/docs/features/classifier-router#defining-which-model-for-which-case","content":"The router resolves a model using the first of these that applies:\n\n| #   | Mechanism              | How you define it                                                                  | Best for                              |\n| --- | ---------------------- | ---------------------------------------------------------------------------------- | ------------------------------------- |\n| 1   | LLM direct pick    | classifier: \"llm\" + a description on each pool member                          | Custom/registry-less models; smartest |\n| 2   | tierMap          | Explicit difficulty → members map                                                | Full, deterministic control           |\n| 3   | Per-member tiers | tiers: [\"hard\",\"expert\"] on a member                                             | Simple, explicit, generic             |\n| 4   | Metadata scoring   | cost / quality per member (declared, or auto-enriched from the model registry) | Known models with comparable metadata |\n\nWhen two members can't be separated (e.g. equal declared quality, or the model registry only knows both as \"high\" quality), the router keeps the declared pool order. For reliable hard-vs-easy separation, prefer mechanisms 1–3, or give members distinct quality values.","hierarchy":{"lvl0":"Features","lvl1":"Classifier Router","lvl2":"Defining \"which model for which case\"","lvl3":""}},
 {"objectID":"bd4863bf711f00f17fb26ea1e60a541aef1f396afc1050ee47b9e29731c90f9a","title":"Metadata scoring rules","url":"/docs/features/classifier-router#metadata-scoring-rules","content":"trivial / simple → cheapest first (cost ascending)\nmoderate → best quality − cost\nhard / expert → most capable first (quality descending)\n\nMembers may declare cost (relative, lower = cheaper) and quality (relative, higher = more capable). If omitted, NeuroLink tries to enrich them from its model registry; if the model is unknown (e.g. a custom LiteLLM endpoint), use mechanisms 1–3 instead.","hierarchy":{"lvl0":"Features","lvl1":"Classifier Router","lvl2":"Metadata scoring rules","lvl3":""}},
 {"objectID":"9e5f93906c1460117b27372537b30d578daec8b1d3fc38ecf1b0fb05c9944801","title":"Custom & self-hosted models (LiteLLM, OpenAI-compatible, Ollama)","url":"/docs/features/classifier-router#custom-self-hosted-models-litellm-openai-compatible-ollama","content":"These models aren't in any registry, so define routing explicitly — both approaches are fully generic:\n\nHeuristic + tiers (deterministic, no LLM cost):\n\nLLM picks per-prompt from plain-English descriptions (most flexible):\n\nThe classifier model is shown each candidate's id (defaults to provider/model) and description, and returns the best id for the prompt. An invalid or absent pick falls back to difficulty-based selection; any classifier failure falls back to the heuristic.","hierarchy":{"lvl0":"Features","lvl1":"Classifier Router","lvl2":"Custom & self-hosted models (LiteLLM, OpenAI-compatible, Ollama)","lvl3":""}},
@@ -3246,7 +3247,7 @@
 {"objectID":"a18da32e2f63d235059c36cf45cff486c8f1dc0c4d4e55ee5f239dd89ede5d21","title":"BudgetCheckResult","url":"/docs/features/context-compaction#budgetcheckresult","content":"Returned by checkContextBudget().","hierarchy":{"lvl0":"Features","lvl1":"Context Compaction","lvl2":"BudgetCheckResult","lvl3":""}},
 {"objectID":"1a2b7ee1dace651a8300809eec11e00b29da084cbfe5b040e4760d58ea0b5879","title":"BudgetCheckParams","url":"/docs/features/context-compaction#budgetcheckparams","content":"Parameters for checkContextBudget().\n\nSource: src/lib/context/budgetChecker.ts:18-54","hierarchy":{"lvl0":"Features","lvl1":"Context Compaction","lvl2":"BudgetCheckParams","lvl3":""}},
 {"objectID":"4bad23fd2858cc71b5af8506276229485c7d0a9636cfddae5c5d9df400bc28b8","title":"The 5-Stage Pipeline","url":"/docs/features/context-compaction#the-5-stage-pipeline","content":"The ContextCompactor runs stages sequentially. Each stage only runs if the\nprevious stage didn't bring tokens below the target budget.\n\n| #   | Stage                     | CompactionStage | Needs a decision model                 |\n| --- | ------------------------- | ----------------- | -------------------------------------- |\n| 0   | Relevance drop            | relevance       | yes — skipped entirely without one |\n| 1   | Tool output pruning       | prune           | no                                     |\n| 2   | File read deduplication   | deduplicate     | no                                     |\n| 3   | LLM summarization         | summarize       | no (its gate uses one)               |\n| 4   | Sliding window truncation | truncate        | no                                     |\n\nresult.stagesUsed reports the ones that actually ran, in order.","hierarchy":{"lvl0":"Features","lvl1":"Context Compaction","lvl2":"The 5-Stage Pipeline","lvl3":""}},
-{"objectID":"918fce938e3acd086621a86be0faaaecccf78059d8a0ef53b2dbc4fcd87347a1","title":"Stage 0: Relevance drop","url":"/docs/features/context-compaction#stage-0-relevance-drop","content":"File: src/lib/context/contextDecision.ts\n\nEverything below Stage 0 is chronological: the pipeline's only notion of\n\"droppable\" is \"old\". Stage 0 is the one stage that asks what a message is\nfor — one boolean per message (\"is this needed to answer the current\nrequest?\") in a single batch. On TypeSafe that costs the same for 200 messages\nas for one because decision latency is flat in question count; on Perplexity each\nfurther question adds about 65 ms and a request takes at most 128 (see\nits guide); on\nCloudflare Clef a request takes at most 64 questions, which tryDecide() splits\ninto batches of 64 (see its guide);\nOpenAI Decisions declares no question cap, so the request goes whole and what the\nendpoint accepts is not established\n(see its guide).\nNeuroLink also refuses a state it estimates at more than 16,000 tokens (about\n60,000 characters of prose), so a longer set of eligible messages gets a refusal\nand the stage is skipped, as on any failure.\n\nIt is strictly additive. With no decision provider configured the stage\ndoes not run, stagesUsed omits relevance, and the pipeline behaves exactly\nas the four-stage one always did. It is also bounded by maxDropRatio and\nwalks oldest-first, so when the cap binds it spares the newest candidates —\nthe same recency assumption every other stage makes.","hierarchy":{"lvl0":"Features","lvl1":"Context Compaction","lvl2":"Stage 0: Relevance drop","lvl3":""}},
+{"objectID":"918fce938e3acd086621a86be0faaaecccf78059d8a0ef53b2dbc4fcd87347a1","title":"Stage 0: Relevance drop","url":"/docs/features/context-compaction#stage-0-relevance-drop","content":"File: src/lib/context/contextDecision.ts\n\nEverything below Stage 0 is chronological: the pipeline's only notion of\n\"droppable\" is \"old\". Stage 0 is the one stage that asks what a message is\nfor — one boolean per message (\"is this needed to answer the current\nrequest?\") in a single batch. On TypeSafe that costs the same for 200 messages\nas for one because decision latency is flat in question count; on Perplexity each\nfurther question adds about 65 ms and a request takes at most 128 (see\nits guide); on\nCloudflare Clef a request takes at most 64 questions, which tryDecide() splits\ninto batches of 64 (see its guide);\nOpenAI Decisions declares no question cap, so the request goes whole and what the\nendpoint accepts is not established\n(see its guide);\nScaleDown declares none either, and its maximum is not measured\n(see its guide).\nThe 16,000-token refusal is Cloudflare Clef's own: NeuroLink refuses a state it\nestimates at more than 16,000 tokens (about 60,000 characters of prose) there, so\na longer set of eligible messages gets a refusal and the stage is skipped, as on\nany failure. A provider that declares no state limit, ScaleDown included, is sent\nthe whole state, which this stage bounds only at 300 messages of at most 1,200\ncharacters each.\n\nIt is strictly additive. With no decision provider configured the stage\ndoes not run, stagesUsed omits relevance, and the pipeline behaves exactly\nas the four-stage one always did. It is also bounded by maxDropRatio and\nwalks oldest-first, so when the cap binds it spares the newest candidates —\nthe same recency assumption every other stage makes.","hierarchy":{"lvl0":"Features","lvl1":"Context Compaction","lvl2":"Stage 0: Relevance drop","lvl3":""}},
 {"objectID":"c1fa9b4e362150149aa7889fbd1374625457cdc5cba6f57fb15d82ef305fc492","title":"The summary gate","url":"/docs/features/context-compaction#the-summary-gate","content":"Stage 3 used to accept any non-empty string as a summary. When a decision\nmodel is configured, the generated summary is now checked first (\"does this\npreserve every decision and open question?\") and a rejected summary leaves the\nmessages untouched so a later stage can try instead. The rejection is recorded\non the span as compaction.stage3.summaryRejected, because a gate that\nsilently discarded work would be indistinguishable from one that never ran.\n\nRejection is deliberately rare: the gate exists to catch a summary that lost a\ndecision, not to second-guess wording.","hierarchy":{"lvl0":"Features","lvl1":"Context Compaction","lvl2":"The summary gate","lvl3":""}},
 {"objectID":"2a518830a84553879be50318dfc47b77a34bf2c61cb5bc21a443ad5a5fcba95b","title":"Stage 1: Tool Output Pruning","url":"/docs/features/context-compaction#stage-1-tool-output-pruning","content":"File: src/lib/context/stages/toolOutputPruner.ts\n\nWalks messages backwards, protecting the most recent tool outputs, and replaces older tool results with \"[Tool result cleared]\".\n\nPruneConfig:\n\n| Field            | Type       | Default     | Description                                                 |\n| ---------------- | ---------- | ----------- | ----------------------------------------------------------- |\n| protectTokens  | number   | 40,000    | Token budget of recent tool outputs to protect from pruning |\n| minimumSavings | number   | 20,000    | Minimum tokens that must be saved for pruning to be applied |\n| protectedTools | string[] | [\"skill\"] | Tool names that are never pruned                            |\n| provider       | string   | —           | Provider name for token estimation multiplier               |\n\nPruneResult:","hierarchy":{"lvl0":"Features","lvl1":"Context Compaction","lvl2":"Stage 1: Tool Output Pruning","lvl3":""}},
 {"objectID":"7c30f222482767920767c8a3dee1c3955d4b615a31c54b58851b467e409a9e64","title":"Stage 2: File Read Deduplication","url":"/docs/features/context-compaction#stage-2-file-read-deduplication","content":"File: src/lib/context/stages/fileReadDeduplicator.ts\n\nDetects multiple reads of the same file path. Keeps only the latest read, replaces earlier reads with \"[File <path> - refer to latest read below]\".\n\nDeduplicationResult:\n\nFile read detection uses the regex pattern: /(?:read|reading|read_file|readFile|Read file|cat)\\s+['\"]?([^\\s'\"\\n]+)/i\n\nA 30% savings threshold (DEDUP_THRESHOLD = 0.3) must be met for deduplication to be applied.","hierarchy":{"lvl0":"Features","lvl1":"Context Compaction","lvl2":"Stage 2: File Read Deduplication","lvl3":""}},
@@ -3357,23 +3358,24 @@
 {"objectID":"a78bcf758e0d2cbf2d2e9ad39bb04c851848618fb55718447904be31d350055b","title":"Limitations","url":"/docs/features/csv-support#limitations","content":"Max file size: 10MB by default (configurable)\nMax rows: 1000 by default (configurable)\nEncoding: auto-detected via BOM + chardet (UTF-8 / UTF-16 / Windows-1252 / Latin-1 …), or forced with encoding (#362)\nPer-row size: a single row is capped at 10MB to bound memory; larger rows fail fast with a clear error (#371)\nParse timeout: parsing is time-bounded (30s strings / 5min files); on timeout partial rows are returned with metadata.parseTimedOut (#379)\nRow shape: parsed rows are validated to be string-keyed objects with string values; a malformed row aborts the parse with [CSVProcessor] Invalid CSV row <n> (#384)\nToken limits: Large CSV files may exceed provider token limits\nStreaming: CSV content is parsed and formatted before sending (not streamed to LLM)","hierarchy":{"lvl0":"Features","lvl1":"CSV File Support","lvl2":"Limitations","lvl3":""}},
 {"objectID":"1e48831d1c4fc775af2633d0340bdeb3d11c96abc213cbdd814a53d14ce500df","title":"Related Features","url":"/docs/features/csv-support#related-features","content":"Office Documents: DOCX, PPTX, XLSX processing\nPDF Support: PDF document processing\nImage Support: Similar multimodal input for images\nFile Detection: Auto-detect file types with confidence scores\nMemory Efficient: Streaming parser for large files\nProvider Agnostic: Works across supported providers\nCLI Integration: Full CLI support with options","hierarchy":{"lvl0":"Features","lvl1":"CSV File Support","lvl2":"Related Features","lvl3":""}},
 {"objectID":"19734b9ffaeb8376dfcebd944730a4cedae77bc1e7ed769733b8b71b1758c421","title":"Summary","url":"/docs/features/csv-support#summary","content":"CSV support is multimodal input (like images)\nUse csvFiles array or files array (auto-detect)\nCustomize with csvOptions (maxRows, formatStyle, includeHeaders)\nWorks across supported providers (not just vision models)\nMemory efficient streaming parser\nCLI support with --csv, --file, --csv-max-rows, --csv-format\nOnly types exposed from package (not classes)","hierarchy":{"lvl0":"Features","lvl1":"CSV File Support","lvl2":"Summary","lvl3":""}},
-{"objectID":"fce0b532226adb069bbeb5459dc5e4ec28037617ca805845768d918653ddbfd5","title":"The `decide` inference type","url":"/docs/features/decide-inference-type","content":"The decide inference type\n\nDeep-dive: generate, stream, decide: a third inference type for NeuroLink —\nwhy this shipped as a provider rather than a subsystem, the five call sites, and the four\ntransport bugs found by adding the Vercel AI Gateway (two of which produced plausible output).\n\nNeuroLink recognises three inference types. Two of them produce text:\n\n| Type         | Call                     | Produces                       |\n| ------------ | ------------------------ | ------------------------------ |\n| generate   | neurolink.generate()   | text                           |\n| stream     | neurolink.stream()     | text, incrementally            |\n| decide | neurolink.decide() | typed judgements — no text |\n\nA decision model takes one state plus a map of named, typed questions and\nreturns one typed answer per question, all evaluated in a single parallel pass.\nThere is no text anywhere in the response, so nothing has to be parsed back out\nof prose. The decide providers are TypeSafe's Jev, Convai Innovations'\nopen-weights Laya, Juspay's open-weights XOR, which also reads images and\nvideo, Subquadratic's hosted SubQ Snap, Perplexity's hosted Decisions\nAPI, which also reads images,\nCloudflare's Clef models on Workers AI,\nwhich also read images, and OpenAI's\nDecisions API (a public beta), which also reads images.\n\nThis is not neurolink.evaluate(), which scores an\nalready-generated response with RAGAS scorers. Different feature, different\nword.\n\nThe three primitives\n\n| Type      | Question                       | Answer fields                                    |\n| --------- | ------------------------------ | ------------------------------------------------ |\n| boolean | Is this statement true?        | probability (0–1) — no confidence          |\n| choice  | Which option from this set?    | choice, probabilities, confidence          |\n| score   | Rate against an ordered rubric | score, legend, probabilities, confidence |\n\nAll three mix freely in one call.\n\nVocabulary note. TypeSafe calls the yes/no primitive a noul and answers\nit in a field of the same name. The Vercel AI SDK and Pydantic AI both renamed\nthat to boolean/probability when exposing it, and NeuroLink follows them —\nthe vendor's spelling is translated inside TypeSafeProvider, so another\ndecision provider slots in without changing any call site.\n\nConfidence is not probability\n\nprobabilities says what the model thinks. confidence says _whether you\nshould act on it_. On TypeSafe's Jev it is calibrated — derived from the\ndistribution, not self-reported — which is what makes it usable as a gate. Each\nprovider reports its own: Perplexity's is, in Perplexity's words, the model's own\ncertainty estimate and not the top probability, Perplexity does not call it\ncalibrated, and NeuroLink has not measured whether it is, so tune a threshold on\nyour own data (see the\nPerplexity guide).\n\nCalibration is a property of groups of answers, not a promise about any\none. Where a confidence is calibrated, answers scored 0.8 are right about 80%\nof the time across many answers. It does not mean a specific 0.8 answer is right.\n\nA boolean carries no confidence of its own. Use decisionBooleanConfidence(p)\n— distance from a coin flip, so 0.5 → 0 and 0/1 → 1. Note also that a boolean\nand an equivalent two-option choice are not guaranteed to agree, and\ncomplementary booleans do not reliably sum to 1, so a threshold tuned on one\nquestion shape does not transfer to another.\n\nEnabling it\n\nGet a key at console.typesafe.ai/keys.\n\nLaya, at a Laya server or a LiteLLM proxy route\n\nLaya has no built-in endpoint. The base URL and key can equally come from the\nconfig passed to the SDK, new NeuroLink({ credentials: { laya: { baseURL, apiKey } } }),\nor per call; config set there counts when NeuroLink picks the default decision\nprovider, exactly as the environment does.\n\nWhen a TypeSafe key (TYPESAFE_API_KEY or AI_GATEWAY_API_KEY) is configured\nalongside Laya, TypeSafe is the default; Laya runs where a caller names it\n(provider: \"laya\") or when neither TypeSafe key is set. Laya counts as\nconfigured only with both its key and its base URL. See the\nLaya provider guide.\n\nXOR, at a deployment or a LiteLLM proxy route\n\nXOR, Juspay's open-weights decision model, has no built-in endpoint either.\nNeuroLink calls <base URL>/v1/systemone, and a trailing /v1 on the base URL\nis accepted. The base URL must be https://; http:// is accepted only for\nlocalhost, 127.0.0.1 and [::1], because every request carries the key and\nthe state. A redirect from the endpoint is refused, not followed. The base URL\nand key can equally come from the config passed to the SDK,\nnew NeuroLink({ credentials: { xor: { baseURL, apiKey } } }), or per call; config set there counts when NeuroLink picks the default decision\nprovider, exactly as the environment does. On a LiteLLM proxy the key's team\nmust allow xor-1.1, otherwise the proxy answers 403 team_model_access_denied.\n\nBuilt-in features use the first configured decision provid","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"","lvl3":""}},
-{"objectID":"f91e7ef7473aa411ff8c0d3b04381325781477c1876408ffd5469dbaf2d9584c","title":"The decide inference type","url":"/docs/features/decide-inference-type#the-decide-inference-type","content":"Deep-dive: generate, stream, decide: a third inference type for NeuroLink —\nwhy this shipped as a provider rather than a subsystem, the five call sites, and the four\ntransport bugs found by adding the Vercel AI Gateway (two of which produced plausible output).\n\nNeuroLink recognises three inference types. Two of them produce text:\n\n| Type         | Call                     | Produces                       |\n| ------------ | ------------------------ | ------------------------------ |\n| generate   | neurolink.generate()   | text                           |\n| stream     | neurolink.stream()     | text, incrementally            |\n| decide | neurolink.decide() | typed judgements — no text |\n\nA decision model takes one state plus a map of named, typed questions and\nreturns one typed answer per question, all evaluated in a single parallel pass.\nThere is no text anywhere in the response, so nothing has to be parsed back out\nof prose. The decide providers are TypeSafe's Jev, Convai Innovations'\nopen-weights Laya, Juspay's open-weights XOR, which also reads images and\nvideo, Subquadratic's hosted SubQ Snap, Perplexity's hosted Decisions\nAPI, which also reads images,\nCloudflare's Clef models on Workers AI,\nwhich also read images, and OpenAI's\nDecisions API (a public beta), which also reads images.\n\nThis is not neurolink.evaluate(), which scores an\nalready-generated response with RAGAS scorers. Different feature, different\nword.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"The decide inference type","lvl3":""}},
+{"objectID":"fce0b532226adb069bbeb5459dc5e4ec28037617ca805845768d918653ddbfd5","title":"The `decide` inference type","url":"/docs/features/decide-inference-type","content":"The decide inference type\n\nDeep-dive: generate, stream, decide: a third inference type for NeuroLink —\nwhy this shipped as a provider rather than a subsystem, the five call sites, and the four\ntransport bugs found by adding the Vercel AI Gateway (two of which produced plausible output).\n\nNeuroLink recognises three inference types. Two of them produce text:\n\n| Type         | Call                     | Produces                       |\n| ------------ | ------------------------ | ------------------------------ |\n| generate   | neurolink.generate()   | text                           |\n| stream     | neurolink.stream()     | text, incrementally            |\n| decide | neurolink.decide() | typed judgements — no text |\n\nA decision model takes one state plus a map of named, typed questions and\nreturns one typed answer per question, all evaluated in a single parallel pass.\nThere is no text anywhere in the response, so nothing has to be parsed back out\nof prose. The decide providers are TypeSafe's Jev, Convai Innovations'\nopen-weights Laya, Juspay's open-weights XOR, which also reads images and\nvideo, Subquadratic's hosted SubQ Snap, Perplexity's hosted Decisions\nAPI, which also reads images,\nCloudflare's Clef models on Workers AI,\nwhich also read images, OpenAI's\nDecisions API (a public beta), which also reads images, and ScaleDown's\nhosted Classify models, which read text only.\n\nThis is not neurolink.evaluate(), which scores an\nalready-generated response with RAGAS scorers. Different feature, different\nword.\n\nThe three primitives\n\n| Type      | Question                       | Answer fields                                    |\n| --------- | ------------------------------ | ------------------------------------------------ |\n| boolean | Is this statement true?        | probability (0–1) — no confidence          |\n| choice  | Which option from this set?    | choice, probabilities, confidence          |\n| score   | Rate against an ordered rubric | score, legend, probabilities, confidence |\n\nAll three mix freely in one call.\n\nVocabulary note. TypeSafe calls the yes/no primitive a noul and answers\nit in a field of the same name. The Vercel AI SDK and Pydantic AI both renamed\nthat to boolean/probability when exposing it, and NeuroLink follows them —\nthe vendor's spelling is translated inside TypeSafeProvider, so another\ndecision provider slots in without changing any call site.\n\nConfidence is not probability\n\nprobabilities says what the model thinks. confidence says _whether you\nshould act on it_. On TypeSafe's Jev it is calibrated — derived from the\ndistribution, not self-reported — which is what makes it usable as a gate. Each\nprovider reports its own: Perplexity's is, in Perplexity's words, the model's own\ncertainty estimate and not the top probability, Perplexity does not call it\ncalibrated, and NeuroLink has not measured whether it is, so tune a threshold on\nyour own data (see the\nPerplexity guide).\n\nCalibration is a property of groups of answers, not a promise about any\none. Where a confidence is calibrated, answers scored 0.8 are right about 80%\nof the time across many answers. It does not mean a specific 0.8 answer is right.\n\nA boolean carries no confidence of its own. Use decisionBooleanConfidence(p)\n— distance from a coin flip, so 0.5 → 0 and 0/1 → 1. Note also that a boolean\nand an equivalent two-option choice are not guaranteed to agree, and\ncomplementary booleans do not reliably sum to 1, so a threshold tuned on one\nquestion shape does not transfer to another.\n\nEnabling it\n\nGet a key at console.typesafe.ai/keys.\n\nLaya, at a Laya server or a LiteLLM proxy route\n\nLaya has no built-in endpoint. The base URL and key can equally come from the\nconfig passed to the SDK, new NeuroLink({ credentials: { laya: { baseURL, apiKey } } }),\nor per call; config set there counts when NeuroLink picks the default decision\nprovider, exactly as the environment does.\n\nWhen a TypeSafe key (TYPESAFE_API_KEY or AI_GATEWAY_API_KEY) is configured\nalongside Laya, TypeSafe is the default; Laya runs where a caller names it\n(provider: \"laya\") or when neither TypeSafe key is set. Laya counts as\nconfigured only with both its key and its base URL. See the\nLaya provider guide.\n\nXOR, at a deployment or a LiteLLM proxy route\n\nXOR, Juspay's open-weights decision model, has no built-in endpoint either.\nNeuroLink calls <base URL>/v1/systemone, and a trailing /v1 on the base URL\nis accepted. The base URL must be https://; http:// is accepted only for\nlocalhost, 127.0.0.1 and [::1], because every request carries the key and\nthe state. A redirect from the endpoint is refused, not followed. The base URL\nand key can equally come from the config passed to the SDK,\nnew NeuroLink({ credentials: { xor: { baseURL, apiKey } } }), or per call; config set there counts when NeuroLink picks the default decision\nprovider, exactly as the environment does. On a LiteLLM proxy the key's team\nmust allow xor-1.1, otherwise the proxy answers 403 team_model_access_denied.\n\n","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"","lvl3":""}},
+{"objectID":"f91e7ef7473aa411ff8c0d3b04381325781477c1876408ffd5469dbaf2d9584c","title":"The decide inference type","url":"/docs/features/decide-inference-type#the-decide-inference-type","content":"Deep-dive: generate, stream, decide: a third inference type for NeuroLink —\nwhy this shipped as a provider rather than a subsystem, the five call sites, and the four\ntransport bugs found by adding the Vercel AI Gateway (two of which produced plausible output).\n\nNeuroLink recognises three inference types. Two of them produce text:\n\n| Type         | Call                     | Produces                       |\n| ------------ | ------------------------ | ------------------------------ |\n| generate   | neurolink.generate()   | text                           |\n| stream     | neurolink.stream()     | text, incrementally            |\n| decide | neurolink.decide() | typed judgements — no text |\n\nA decision model takes one state plus a map of named, typed questions and\nreturns one typed answer per question, all evaluated in a single parallel pass.\nThere is no text anywhere in the response, so nothing has to be parsed back out\nof prose. The decide providers are TypeSafe's Jev, Convai Innovations'\nopen-weights Laya, Juspay's open-weights XOR, which also reads images and\nvideo, Subquadratic's hosted SubQ Snap, Perplexity's hosted Decisions\nAPI, which also reads images,\nCloudflare's Clef models on Workers AI,\nwhich also read images, OpenAI's\nDecisions API (a public beta), which also reads images, and ScaleDown's\nhosted Classify models, which read text only.\n\nThis is not neurolink.evaluate(), which scores an\nalready-generated response with RAGAS scorers. Different feature, different\nword.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"The decide inference type","lvl3":""}},
 {"objectID":"2c0f154b9ae29f0af1a097de9ccb435db1e8f5c35a4df8bb91526eadb7b5eb97","title":"The three primitives","url":"/docs/features/decide-inference-type#the-three-primitives","content":"| Type      | Question                       | Answer fields                                    |\n| --------- | ------------------------------ | ------------------------------------------------ |\n| boolean | Is this statement true?        | probability (0–1) — no confidence          |\n| choice  | Which option from this set?    | choice, probabilities, confidence          |\n| score   | Rate against an ordered rubric | score, legend, probabilities, confidence |\n\nAll three mix freely in one call.\n\nVocabulary note. TypeSafe calls the yes/no primitive a noul and answers\nit in a field of the same name. The Vercel AI SDK and Pydantic AI both renamed\nthat to boolean/probability when exposing it, and NeuroLink follows them —\nthe vendor's spelling is translated inside TypeSafeProvider, so another\ndecision provider slots in without changing any call site.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"The three primitives","lvl3":""}},
 {"objectID":"ff28050e2ddf3c21c910bc84ca32105968e1f7ffdbc3b19a983bb9adc6e9cc36","title":"Confidence is not probability","url":"/docs/features/decide-inference-type#confidence-is-not-probability","content":"probabilities says what the model thinks. confidence says _whether you\nshould act on it_. On TypeSafe's Jev it is calibrated — derived from the\ndistribution, not self-reported — which is what makes it usable as a gate. Each\nprovider reports its own: Perplexity's is, in Perplexity's words, the model's own\ncertainty estimate and not the top probability, Perplexity does not call it\ncalibrated, and NeuroLink has not measured whether it is, so tune a threshold on\nyour own data (see the\nPerplexity guide).\n\nCalibration is a property of groups of answers, not a promise about any\none. Where a confidence is calibrated, answers scored 0.8 are right about 80%\nof the time across many answers. It does not mean a specific 0.8 answer is right.\n\nA boolean carries no confidence of its own. Use decisionBooleanConfidence(p)\n— distance from a coin flip, so 0.5 → 0 and 0/1 → 1. Note also that a boolean\nand an equivalent two-option choice are not guaranteed to agree, and\ncomplementary booleans do not reliably sum to 1, so a threshold tuned on one\nquestion shape does not transfer to another.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Confidence is not probability","lvl3":""}},
 {"objectID":"91bfa1275464235c40e00d8a9ac0805e9e181450cccb6235fabd43178bef7c02","title":"Enabling it","url":"/docs/features/decide-inference-type#enabling-it","content":"Get a key at console.typesafe.ai/keys.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Enabling it","lvl3":""}},
 {"objectID":"71509585343bb3860ff000ad67405f6ed19fbc26f8eb973ac1fe58088af460b4","title":"Laya, at a Laya server or a LiteLLM proxy route","url":"/docs/features/decide-inference-type#laya-at-a-laya-server-or-a-litellm-proxy-route","content":"Laya has no built-in endpoint. The base URL and key can equally come from the\nconfig passed to the SDK, new NeuroLink({ credentials: { laya: { baseURL, apiKey } } }),\nor per call; config set there counts when NeuroLink picks the default decision\nprovider, exactly as the environment does.\n\nWhen a TypeSafe key (TYPESAFE_API_KEY or AI_GATEWAY_API_KEY) is configured\nalongside Laya, TypeSafe is the default; Laya runs where a caller names it\n(provider: \"laya\") or when neither TypeSafe key is set. Laya counts as\nconfigured only with both its key and its base URL. See the\nLaya provider guide.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Laya, at a Laya server or a LiteLLM proxy route","lvl3":""}},
-{"objectID":"19994573e52cd09183db8050eadaea8ece3e291ef40c3009a5c44dfa959bf99c","title":"XOR, at a deployment or a LiteLLM proxy route","url":"/docs/features/decide-inference-type#xor-at-a-deployment-or-a-litellm-proxy-route","content":"XOR, Juspay's open-weights decision model, has no built-in endpoint either.\nNeuroLink calls <base URL>/v1/systemone, and a trailing /v1 on the base URL\nis accepted. The base URL must be https://; http:// is accepted only for\nlocalhost, 127.0.0.1 and [::1], because every request carries the key and\nthe state. A redirect from the endpoint is refused, not followed. The base URL\nand key can equally come from the config passed to the SDK,\nnew NeuroLink({ credentials: { xor: { baseURL, apiKey } } }), or per call; config set there counts when NeuroLink picks the default decision\nprovider, exactly as the environment does. On a LiteLLM proxy the key's team\nmust allow xor-1.1, otherwise the proxy answers 403 team_model_access_denied.\n\nBuilt-in features use the first configured decision provider in the order\nTypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef, OpenAI Decisions. XOR counts as configured only with both its key\nand its base URL, and runs where a caller names it (provider: \"xor\") or when\nneither TypeSafe nor Laya is configured. It reads images and video; see\nImages and video and the\nXOR provider guide.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"XOR, at a deployment or a LiteLLM proxy route","lvl3":""}},
+{"objectID":"19994573e52cd09183db8050eadaea8ece3e291ef40c3009a5c44dfa959bf99c","title":"XOR, at a deployment or a LiteLLM proxy route","url":"/docs/features/decide-inference-type#xor-at-a-deployment-or-a-litellm-proxy-route","content":"XOR, Juspay's open-weights decision model, has no built-in endpoint either.\nNeuroLink calls <base URL>/v1/systemone, and a trailing /v1 on the base URL\nis accepted. The base URL must be https://; http:// is accepted only for\nlocalhost, 127.0.0.1 and [::1], because every request carries the key and\nthe state. A redirect from the endpoint is refused, not followed. The base URL\nand key can equally come from the config passed to the SDK,\nnew NeuroLink({ credentials: { xor: { baseURL, apiKey } } }), or per call; config set there counts when NeuroLink picks the default decision\nprovider, exactly as the environment does. On a LiteLLM proxy the key's team\nmust allow xor-1.1, otherwise the proxy answers 403 team_model_access_denied.\n\nBuilt-in features use the first configured decision provider in the order\nTypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef, OpenAI Decisions, ScaleDown. XOR counts as configured only with both its key\nand its base URL, and runs where a caller names it (provider: \"xor\") or when\nneither TypeSafe nor Laya is configured. It reads images and video; see\nImages and video and the\nXOR provider guide.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"XOR, at a deployment or a LiteLLM proxy route","lvl3":""}},
 {"objectID":"99d5ee10748044f10d6dd1e54785276aac1cf4375eca8dc6a1c9db90fc0c2158","title":"SubQ, at its hosted endpoint","url":"/docs/features/decide-inference-type#subq-at-its-hosted-endpoint","content":"Subquadratic's subq-snap-preview (public preview) speaks the same System One\nwire as TypeSafe, at a public endpoint, so a key alone configures it and there\nis no base URL to set. The key and base URL can equally come from the config\npassed to the SDK, new NeuroLink({ credentials: { subq: { apiKey, baseURL } } }),\nor per call. The provider id is subq; subquadratic and subq-snap name it\ntoo. It comes after TypeSafe, Laya and XOR and ahead of Perplexity and Cloudflare\nClef: its key belongs to no other provider, so it is set on purpose, whereas\ntheirs are shared with text providers and must not displace it. NeuroLink sends it no images or video: SubQ's image format is not\none the decide catalog can describe yet, so an image attached to a call is\nrefused before any request. The state window NeuroLink reports for it is advisory\n(reported, never enforced, not a guarantee); see the\nSubQ provider guide.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"SubQ, at its hosted endpoint","lvl3":""}},
 {"objectID":"dc8b9e9a2c66b1d2a60305a09a44263ceca755d40f04b6771e0a7c05a5237a64","title":"Perplexity, at its hosted endpoint","url":"/docs/features/decide-inference-type#perplexity-at-its-hosted-endpoint","content":"Perplexity's pplx-decider-v1-27b is a hosted API at a public endpoint, so a key\nalone configures it and there is no base URL to set. NeuroLink calls\n<base URL>/v1/decisions; a trailing /v1 on an override is accepted. The key\ncan equally come from the config passed to the SDK,\nnew NeuroLink({ credentials: { perplexityDecider: { apiKey } } }), or per call;\nconfig set there counts when a decide() call picks the default decision\nprovider, exactly as the environment does. The classifier router's auto gate\ncounts them too, per-call credentials included, so a\ncredentials.perplexityDecider passed on one call can make auto pick jev.\nThe provider id is perplexity-decider: the perplexity provider is the\nSonar text provider, which serves generate() and stream() and not decide().\n\nPERPLEXITY_API_KEY is shared with that text provider. A host that set it\nonly to use Sonar has therefore also configured decide, and built-in features\nwill use Perplexity whenever none of TypeSafe, Laya, XOR or SubQ is configured.\nPerplexity comes after XOR and SubQ in descriptor order, so it never displaces\none that is. Context compaction's relevance stage and summary gate need no opt-in of their\nown, so earlier conversation text starts going to Perplexity the first time a\nconversation outgrows its budget; model routing, tool routing and RAG planning do\nnothing unless enabled. To keep the shared key from activating decisions, pass\nthe text provider's key as credentials.perplexity instead of through the\nenvironment, and keep it out of .env too, because the SDK and the CLI load that\nfile into the environment, or configure one of TypeSafe, Laya, XOR and SubQ, which then\nreceives those texts instead. The Perplexity provider guide lists\nwhat each consumer sends\nand the\nexact switches.\nIt reads images but no video; see Images and video.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Perplexity, at its hosted endpoint","lvl3":""}},
 {"objectID":"86d184ec9e501d4e94e973c833eaee71bb40744e683334ad36c497c3f625387f","title":"Cloudflare Clef, on Workers AI","url":"/docs/features/decide-inference-type#cloudflare-clef-on-workers-ai","content":"Cloudflare's clef and clef-flash are hosted on Workers AI at a public\nendpoint, so there is no base URL to set. The account id is part of the route, so\nClef counts as configured only with both the token and the account id. They can\nequally come from the config passed to the SDK,\nnew NeuroLink({ credentials: { cloudflareClef: { apiKey, accountId } } })\n(baseURL is optional), or per call; config set there counts when a decide()\ncall picks the default decision provider, exactly as the environment does. The\nclassifier router's auto gate counts them too, per-call credentials included,\nso a credentials.cloudflareClef passed on one call can make auto pick jev.\nThe provider id is cloudflare-clef: the cloudflare provider is the Workers AI text\nprovider, which serves generate() and stream() and not decide(), and\ncredentials.cloudflare does not configure decide.\n\nThe Workers AI endpoint's state window changed during October 2026.\nCloudflare documents a 64K window; until 2026-10-04 text past about 2,048 tokens\nwas ignored without an error, and on 2026-10-07 states of at least 190,153 tokens\n(clef-flash) and 52,154 (clef) were read in full (hosted service or model:\nunknown). NeuroLink refuses a state it estimates at more than 16,000 tokens with\nmax_tokens_exceeded, which every built-in consumer treats as \"carry on as\nbefore\"; see Limits and gotchas. Latency grows with the\nstate, about 0.2 s per 1,000 tokens on large ones, so NeuroLink's default timeout\nadds 250 ms per 1,000 estimated tokens.\n\nCLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID are shared with that text\nprovider. A host that set them only to use Workers AI text has therefore also\nconfigured decide, and built-in features will use Clef whenever none of\nTypeSafe, Laya, XOR, SubQ or Perplexity is configured. Clef comes after\nTypeSafe, Laya, XOR, SubQ and Perplexity, so it never displaces one that is. No switch turns it off while the two\nvariables are set.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Cloudflare Clef, on Workers AI","lvl3":""}},
-{"objectID":"08fd87a8baa41fd742824c8457907e72804c393825cb2f67997a356c98b4211b","title":"OpenAI Decisions, at its hosted endpoint","url":"/docs/features/decide-inference-type#openai-decisions-at-its-hosted-endpoint","content":"OpenAI's Decisions API is a public beta at a public endpoint, so a key alone\nconfigures it and there is no base URL to set. NeuroLink calls\n<base URL>/v1/decisions; a trailing /v1 on an override is accepted. The key\ncan equally come from the config passed to the SDK,\nnew NeuroLink({ credentials: { openaiDecisions: { apiKey } } }), or per call;\nconfig set there counts when a decide() call picks the default decision\nprovider, exactly as the environment does. The classifier router's auto gate is\nread differently: it looks at the constructor's credentials and the environment\nonly, so a credentials.openaiDecisions passed on one call does not influence\nit. The provider id is openai-decisions: the openai provider is the chat\nprovider, which serves generate() and stream() and not decide(), and\ncredentials.openai does not configure decide. Its wire is OpenAI's own (named\nquestions in an array, an array of answers), so NeuroLink translates both ways;\nthe question types you write are unchanged. See the provider guide's\nwire mapping.\n\nOPENAI_API_KEY is shared with that chat provider. A host that set it only to\nuse OpenAI chat has therefore also configured decide, and built-in features will\nuse OpenAI Decisions whenever none of TypeSafe, Laya, XOR, SubQ, Perplexity or Cloudflare\nClef is configured. It comes last in the order, so it never displaces one that is.\nThe built-in consumers (model routing, tool routing, relevance compaction, the\nsummary gate, RAG planning) and the speech-to-text correction check then make paid\ndecision calls, though they still fail open. No switch turns it off while the key\nis set (short of a gateway OPENAI_BASE_URL, below). neurolink.decide() and tryDecide() build a fresh provider for every\ncall, so a key OpenAI rejects costs one failing request per decision site each\ntime, and the 401 latch never carries across calls. The provider guide explains\nwhen NeuroLink uses it\nand\nhow the two providers share the key.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"OpenAI Decisions, at its hosted endpoint","lvl3":""}},
+{"objectID":"08fd87a8baa41fd742824c8457907e72804c393825cb2f67997a356c98b4211b","title":"OpenAI Decisions, at its hosted endpoint","url":"/docs/features/decide-inference-type#openai-decisions-at-its-hosted-endpoint","content":"OpenAI's Decisions API is a public beta at a public endpoint, so a key alone\nconfigures it and there is no base URL to set. NeuroLink calls\n<base URL>/v1/decisions; a trailing /v1 on an override is accepted. The key\ncan equally come from the config passed to the SDK,\nnew NeuroLink({ credentials: { openaiDecisions: { apiKey } } }), or per call;\nconfig set there counts when a decide() call picks the default decision\nprovider, exactly as the environment does. The classifier router's auto gate is\nread differently: it looks at the constructor's credentials and the environment\nonly, so a credentials.openaiDecisions passed on one call does not influence\nit. The provider id is openai-decisions: the openai provider is the chat\nprovider, which serves generate() and stream() and not decide(), and\ncredentials.openai does not configure decide. Its wire is OpenAI's own (named\nquestions in an array, an array of answers), so NeuroLink translates both ways;\nthe question types you write are unchanged. See the provider guide's\nwire mapping.\n\nOPENAI_API_KEY is shared with that chat provider. A host that set it only to\nuse OpenAI chat has therefore also configured decide, and built-in features will\nuse OpenAI Decisions whenever none of TypeSafe, Laya, XOR, SubQ, Perplexity or Cloudflare\nClef is configured. It comes after those six in the order, so it never displaces\none that is; only ScaleDown follows it.\nThe built-in consumers (model routing, tool routing, relevance compaction, the\nsummary gate, RAG planning) and the speech-to-text correction check then make paid\ndecision calls, though they still fail open. No switch turns it off while the key\nis set (short of a gateway OPENAI_BASE_URL, below). neurolink.decide() and tryDecide() build a fresh provider for every\ncall, so a key OpenAI rejects costs one failing request per decision site each\ntime, and the 401 latch never carries across calls. The provider guide explains\nwhen NeuroLink uses it\nand\nhow the two providers share the key.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"OpenAI Decisions, at its hosted endpoint","lvl3":""}},
+{"objectID":"691b7e7dec264cbbdcf43939939e82963f66753635c1b5ca81b26b2be05c844b","title":"ScaleDown, at its hosted endpoint","url":"/docs/features/decide-inference-type#scaledown-at-its-hosted-endpoint","content":"ScaleDown's Classify models speak the same System One wire as TypeSafe, at a\npublic endpoint, so a key alone configures it and there is no base URL to set.\nNeuroLink calls <base URL>/v1/scaledown, drops trailing slashes from an\noverride, and sends the key in an x-api-key header and never as a Bearer token.\nThe key can equally come from the config passed to the SDK,\nnew NeuroLink({ credentials: { scaledown: { apiKey } } }), or per call. The\nprovider id is scaledown; scaledown-classify names it too.\n\nScaleDown is the last decision provider, and its key is its own. Nothing\nelse reads SCALEDOWN_API_KEY, so setting it can only be meant for ScaleDown. It\nstill comes last in the order: a host with only that key gets ScaleDown for the\nbuilt-in features, and a host that also has any earlier provider configured\n(including an OPENAI_API_KEY, PERPLEXITY_API_KEY or Cloudflare pair set only for\na text provider) gets that earlier provider and reaches ScaleDown only by naming\nit. When ScaleDown is the default, the built-in consumers make paid decision\ncalls, and still fail open.\n\nIt reads text only, and the state always goes out as a string. A string state\nis sent unchanged, any other value as its JSON text, and a string that is empty or\nonly whitespace is refused before any request, as an invalid_request. {} and\n[] are sent as that text. Images and video are refused before any request. Two\noptional keys in the entry's decide block say this to the shared engine,\nauthHeader: \"x-api-key\" and stateEncoding: \"text\"; an entry that omits them\nbehaves as before. The provider guide has the\nrequest and the\nbilling note:\na choice question is scored once per option, so input tokens grow with options\ntimes state length.\n\nMuch about the endpoint is not measured, including its context size, latency,\nrate limits and the maximum number of options, questions and levels; the guide\nlists what is not measured.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"ScaleDown, at its hosted endpoint","lvl3":""}},
 {"objectID":"7990ffc08aca0afbeae4310cfdcd3758daf52da4d9edd1da90a82b683f4d7d41","title":"Two transports","url":"/docs/features/decide-inference-type#two-transports","content":"The same model is reachable two ways. Which one runs is decided once, in the\nconstructor:\n\n|                     | Direct             | Vercel AI Gateway                             |\n| ------------------- | ------------------ | --------------------------------------------- |\n| Key                 | TYPESAFE_API_KEY | AI_GATEWAY_API_KEY                          |\n| Endpoint            | api.typesafe.ai  | ai-gateway.vercel.sh/v4/ai/evaluation-model |\n| Model named in      | request body       | ai-model-id header                          |\n| Question vocabulary | noul             | boolean                                     |\n| confidence        | on each answer     | on providerMetadata, not on the answer      |\n| Billed by           | TypeSafe           | Vercel                                        |\n\nHolding both keys keeps the direct transport, so the confidence figures a\nhost already sees do not shift underneath it when a second key appears. Both\ntransports report the vendor's calibrated confidence — the gateway simply puts\nit somewhere else, under providerMetadata.typesafe.confidence.<questionId>,\nleaving the answer objects without one. Set TYPESAFE_TRANSPORT=gateway (or\ncredentials.typesafe.transport) to override.\n\n⚠️ Read that field, not the distribution peak. It is tempting to take\nmax(probabilities) when an answer carries no confidence, and on a\nnear-certain answer the two agree. On an uncertain one they do not, and not by a\nlittle: a measured four-way choice returned probabilities\n{alpha 0.16, beta 0.28, gamma 0.33, delta 0.23} — a peak of 0.33 against a\nreported confidence of 0.10. That gap straddles the default\nminUpgradeConfidence of 0.3, so the derived number clears a bar the real one\nfails and a near-random pick gets acted on as a confident one. The peak stays as\nthe fallback when neither source reports a confidence, and it is genuinely a\ndifferent quantity: an even distribution over N options lands near 1/N, not 0.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Two transports","lvl3":""}},
-{"objectID":"e5df7db87d4cb31c67fcce769cf5a3573764cb358ed9ee4b41f7ce86cb3954b7","title":"Using it","url":"/docs/features/decide-inference-type#using-it","content":"readDecisionBoolean / readDecisionChoice / readDecisionScore validate at\nruntime and return undefined for a missing id or a mismatched type, so no\ncall site needs a type assertion.\n\nticketText is attacker-controlled, so the verdict is a hint: keep assignTo\nreversible or put a deterministic check in front of it (see the Untrusted input\nbullet under What it is bad at).\n\nUse decide() instead of tryDecide() when you want the failure to surface;\nit throws a ProviderError whose cause carries a typed kind\n(authentication, rate_limit, max_tokens_exceeded, …).\n\nMore questions than a provider takes. A provider that caps the questions in\none request (Laya and Cloudflare Clef take 64) refuses a longer map from decide() with\nmax_tokens_exceeded. tryDecide(), which every built-in consumer calls,\nsplits the map at the cap instead, keeps the questions in the order given, runs\nup to four batches at a time and joins the answers into one result: usage is\nsummed, the latency is the slowest batch's, and requestId lists the batches'\nids. A batch that fails costs only its own questions, which then have no answer\n(the read* helpers return undefined for them, and every built-in consumer\ncarries on without that decision); tryDecide() returns null only when every\nbatch failed. The state and any media are sent with each batch, so a split\nrequest costs the state's tokens once for each batch. A provider with no cap\n(TypeSafe, XOR, OpenAI Decisions) always gets one request.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Using it","lvl3":""}},
-{"objectID":"7232a490e9e72e01dbe1a12de9470a38ed66ef91aa076b8c8cc745770eb03545","title":"From the CLI","url":"/docs/features/decide-inference-type#from-the-cli","content":"The same primitive is available as neurolink decide [state], which calls\ndecide() (not tryDecide()) and prints one line per answer:\n\nTo ask about an image or a video, add --image <path> (repeatable) or\n--video <path>. XOR reads both, Perplexity, Cloudflare Clef and OpenAI Decisions\nread images only, and TypeSafe, Laya and SubQ read neither:\n\nSee the CLI command reference for the full flag\nlist, including --state-file, --questions-file, --image, --video and\n--format json.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"From the CLI","lvl3":""}},
+{"objectID":"e5df7db87d4cb31c67fcce769cf5a3573764cb358ed9ee4b41f7ce86cb3954b7","title":"Using it","url":"/docs/features/decide-inference-type#using-it","content":"readDecisionBoolean / readDecisionChoice / readDecisionScore validate at\nruntime and return undefined for a missing id or a mismatched type, so no\ncall site needs a type assertion.\n\nticketText is attacker-controlled, so the verdict is a hint: keep assignTo\nreversible or put a deterministic check in front of it (see the Untrusted input\nbullet under What it is bad at).\n\nUse decide() instead of tryDecide() when you want the failure to surface;\nit throws a ProviderError whose cause carries a typed kind\n(authentication, rate_limit, max_tokens_exceeded, …).\n\nMore questions than a provider takes. A provider that caps the questions in\none request (Laya and Cloudflare Clef take 64) refuses a longer map from decide() with\nmax_tokens_exceeded. tryDecide(), which every built-in consumer calls,\nsplits the map at the cap instead, keeps the questions in the order given, runs\nup to four batches at a time and joins the answers into one result: usage is\nsummed, the latency is the slowest batch's, and requestId lists the batches'\nids. A batch that fails costs only its own questions, which then have no answer\n(the read* helpers return undefined for them, and every built-in consumer\ncarries on without that decision); tryDecide() returns null only when every\nbatch failed. The state and any media are sent with each batch, so a split\nrequest costs the state's tokens once for each batch. A provider with no cap\n(TypeSafe, XOR, SubQ, OpenAI Decisions, ScaleDown) always gets one request.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Using it","lvl3":""}},
+{"objectID":"7232a490e9e72e01dbe1a12de9470a38ed66ef91aa076b8c8cc745770eb03545","title":"From the CLI","url":"/docs/features/decide-inference-type#from-the-cli","content":"The same primitive is available as neurolink decide [state], which calls\ndecide() (not tryDecide()) and prints one line per answer:\n\nTo ask about an image or a video, add --image <path> (repeatable) or\n--video <path>. XOR reads both, Perplexity, Cloudflare Clef and OpenAI Decisions\nread images only, and TypeSafe, Laya, SubQ and ScaleDown read neither:\n\nSee the CLI command reference for the full flag\nlist, including --state-file, --questions-file, --image, --video and\n--format json.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"From the CLI","lvl3":""}},
 {"objectID":"3b801c727cb399feb7390c3ad67aba9ea5f77581096bb852b7aafb87b70998aa","title":"A choice answer is also a ranking","url":"/docs/features/decide-inference-type#a-choice-answer-is-also-a-ranking","content":"readDecisionChoice returns ranked — every option sorted by probability,\nhighest first. One choice question over N options therefore ranks all N in a\nsingle request. This is the basis for picking from a large catalogue.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"A choice answer is also a ranking","lvl3":""}},
-{"objectID":"e67682840a5ce9650dba2a1d437b0dfbba456cae60fa3f607127e209bf1ec3e7","title":"Images and video","url":"/docs/features/decide-inference-type#images-and-video","content":"A decision provider whose descriptor declares media limits reads images, and\npossibly one video, alongside state. XOR does, and so do Perplexity, Cloudflare\nClef and OpenAI Decisions, which declare images only and no video; TypeSafe, Laya and SubQ\ndo not. Two optional fields on the request carry them, for decide() and\ntryDecide() alike:\nimages — up to 8 images, the limit XOR, Perplexity and OpenAI Decisions each\n  declare (OpenAI's is NeuroLink's own allowance, not a limit OpenAI documents);\n  Cloudflare Clef takes up to 4.\nvideo — one video, for a provider that declares video (XOR).\n\nEach image and the video takes the same three input forms:\na Buffer;\na local file path;\na data:image/…;base64, or data:video/…;base64, URL.\n\nNeuroLink identifies a Buffer or a file from its bytes, not its extension — PNG,\nJPEG, WebP and GIF images, and MP4, MOV and WebM video —\nand sends it as a data: URL. A data: URL you pass yourself must hold base64\nimage or video content. Clef normalizes image/jpg to image/jpeg; other data URLs are sent as given. Perplexity and Cloudflare Clef read\nPNG, JPEG and WebP only, and refuse a GIF before any request. OpenAI Decisions\nsends every image as an inline data URL and checks no format list, because\nOpenAI's list is not established.\n\nWhat is refused. Everything NeuroLink can check is refused before any\nrequest, as a non-retryable invalid_request:\nan http(s) URL — media is not fetched for you, so pass a Buffer, a file path\n  or a data: URL;\na string that is not a file path or a data: URL, such as bare base64 or a\n  URL with another scheme;\na missing file, a directory, or an empty Buffer or file;\nsomething that is not an image or a video;\nmore than 8 images, or more than 4 for Cloudflare Clef;\na request body over the provider's limit: 8 MB for XOR, 32 MiB for Perplexity\n  and for OpenAI Decisions (NeuroLink's own number for the latter), 256,000 bytes\n  for Cloudflare Clef.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Images and video","lvl3":""}},
-{"objectID":"a6077cedd1ca3e03b757da9a2ef8c185a56d2f1032dd5ed43fe28bd71454836c","title":"The one rule: batch, never fan out","url":"/docs/features/decide-inference-type#the-one-rule-batch-never-fan-out","content":"This inverts the instinct you have from LLMs.\n\nOn TypeSafe, question count barely affects latency (measured against its live\nAPI):\n\n| questions | round trip | input tokens |\n| --------- | ---------- | ------------ |\n| 1         | 393 ms     | 310          |\n| 10        | 390 ms     | 481          |\n| 100       | 423 ms     | 2 281        |\n| 400       | 465 ms     | 8 581        |\n\nThese are TypeSafe's figures. perplexity-decider differs: about 0.3 to 1.1 s\nfor one question, about 65 ms for each further question (a request takes at\nmost 128), and an input time that grows faster than linearly with size: 3.9 s\nat 65,000 tokens, 10 s at 146,000 and 22.9 s at 251,000. See\nits latency section.\ncloudflare-clef takes 0.3 to 1.0 s for a small request and 1.1 to 1.3 s for\n64 questions (measured 2026-10-03). A 64-question request with the same\nquestions and a shorter state took 1.5 s on clef-flash and 2.3 s on clef on 2026-10-04; see\nits latency section.\nopenai-decisions has not been timed: its latency is not measured.\n\nOn TypeSafe, 400 questions cost ~70 ms more than one. Concurrent requests, by\ncontrast, queue: ten parallel calls take ~1.4 s wall with nine landing together\nat the end, while the server's own upstream time stays flat at 64–169 ms.\n\nSo on TypeSafe, 400 things in one request take ~465 ms; the same 400 as separate\nrequests take roughly a minute. Add every question you might need to the call\nyou are already making — speculative questions are nearly free on TypeSafe (on\nPerplexity each costs about 65 ms, up to 128 in a request; Cloudflare Clef takes\nup to 64), and a second round trip is not.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"The one rule: batch, never fan out","lvl3":""}},
+{"objectID":"e67682840a5ce9650dba2a1d437b0dfbba456cae60fa3f607127e209bf1ec3e7","title":"Images and video","url":"/docs/features/decide-inference-type#images-and-video","content":"A decision provider whose descriptor declares media limits reads images, and\npossibly one video, alongside state. XOR does, and so do Perplexity, Cloudflare\nClef and OpenAI Decisions, which declare images only and no video; TypeSafe, Laya, SubQ\nand ScaleDown do not. Two optional fields on the request carry them, for decide() and\ntryDecide() alike:\nimages — up to 8 images, the limit XOR, Perplexity and OpenAI Decisions each\n  declare (OpenAI's is NeuroLink's own allowance, not a limit OpenAI documents);\n  Cloudflare Clef takes up to 4.\nvideo — one video, for a provider that declares video (XOR).\n\nEach image and the video takes the same three input forms:\na Buffer;\na local file path;\na data:image/…;base64, or data:video/…;base64, URL.\n\nNeuroLink identifies a Buffer or a file from its bytes, not its extension — PNG,\nJPEG, WebP and GIF images, and MP4, MOV and WebM video —\nand sends it as a data: URL. A data: URL you pass yourself must hold base64\nimage or video content. Clef normalizes image/jpg to image/jpeg; other data URLs are sent as given. Perplexity and Cloudflare Clef read\nPNG, JPEG and WebP only, and refuse a GIF before any request. OpenAI Decisions\nsends every image as an inline data URL and checks no format list, because\nOpenAI's list is not established.\n\nWhat is refused. Everything NeuroLink can check is refused before any\nrequest, as a non-retryable invalid_request:\nan http(s) URL — media is not fetched for you, so pass a Buffer, a file path\n  or a data: URL;\na string that is not a file path or a data: URL, such as bare base64 or a\n  URL with another scheme;\na missing file, a directory, or an empty Buffer or file;\nsomething that is not an image or a video;\nmore than 8 images, or more than 4 for Cloudflare Clef;\na request body over the provider's limit: 8 MB for XOR, 32 MiB for Perplexity\n  and for OpenAI Decisions (NeuroLink's own number for the latter), 256,000 bytes\n  for Cloudflare Clef.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Images and video","lvl3":""}},
+{"objectID":"a6077cedd1ca3e03b757da9a2ef8c185a56d2f1032dd5ed43fe28bd71454836c","title":"The one rule: batch, never fan out","url":"/docs/features/decide-inference-type#the-one-rule-batch-never-fan-out","content":"This inverts the instinct you have from LLMs.\n\nOn TypeSafe, question count barely affects latency (measured against its live\nAPI):\n\n| questions | round trip | input tokens |\n| --------- | ---------- | ------------ |\n| 1         | 393 ms     | 310          |\n| 10        | 390 ms     | 481          |\n| 100       | 423 ms     | 2 281        |\n| 400       | 465 ms     | 8 581        |\n\nThese are TypeSafe's figures. perplexity-decider differs: about 0.3 to 1.1 s\nfor one question, about 65 ms for each further question (a request takes at\nmost 128), and an input time that grows faster than linearly with size: 3.9 s\nat 65,000 tokens, 10 s at 146,000 and 22.9 s at 251,000. See\nits latency section.\ncloudflare-clef takes 0.3 to 1.0 s for a small request and 1.1 to 1.3 s for\n64 questions (measured 2026-10-03). A 64-question request with the same\nquestions and a shorter state took 1.5 s on clef-flash and 2.3 s on clef on 2026-10-04; see\nits latency section.\nopenai-decisions has not been timed: its latency is not measured. Nor has\nscaledown: its latency is not measured, and a choice question is scored\nonce per option, so a request with many options may need a larger timeoutMs\nthan the 5-second default; see\nits timeouts note.\n\nOn TypeSafe, 400 questions cost ~70 ms more than one. Concurrent requests, by\ncontrast, queue: ten parallel calls take ~1.4 s wall with nine landing together\nat the end, while the server's own upstream time stays flat at 64–169 ms.\n\nSo on TypeSafe, 400 things in one request take ~465 ms; the same 400 as separate\nrequests take roughly a minute. Add every question you might need to the call\nyou are already making — speculative questions are nearly free on TypeSafe (on\nPerplexity each costs about 65 ms, up to 128 in a request; Cloudflare Clef takes\nup to 64), and a second round trip is not.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"The one rule: batch, never fan out","lvl3":""}},
 {"objectID":"0c1ce0b91d624674d6b4cc3b1f67fad986bcdfd379812b89e121784a94408ec3","title":"Model routing","url":"/docs/features/decide-inference-type#model-routing","content":"The classifier router gains a jev\nstrategy, and its default becomes auto — resolving to jev when a decision\nprovider is configured and heuristic when not.\n\nOne request asks for the difficulty tier, whether the task needs\nvision/tools/reasoning, whether carrying it out is risky, and which pool\nmember to use — all at once, in ~400 ms on TypeSafe's Jev.\n\n|                        | heuristic   | llm                           | jev                                               |\n| ---------------------- | ------------- | ------------------------------- | --------------------------------------------------- |\n| Added latency          | 0 ms          | ~1–8 s                          | ~400 ms on TypeSafe                                 |\n| Cost per decision      | none          | a full LLM call                 | ~$0.00002 on TypeSafe                               |\n| Confidence             | keyword score | self-reported (defaults to 0.7) | as the provider reports it (calibrated on TypeSafe) |\n| Picks a model directly | no            | yes                             | yes                                                 |\n\nThe jev column is TypeSafe's figures; the\nPerplexity guide gives its\nlatency and the confidence it reports. The\nCloudflare Clef guide gives its\nlatency; a prompt NeuroLink estimates over 16,000 tokens is refused there and\nrouting falls back to the heuristic, as it does on any failure.\n\nThresholds are asymmetric, because the two mistakes do not cost the same:\nminUpgradeConfidence defaults to 0.3 (spending more on a wrong guess costs\nmoney) and minDowngradeConfidence to 0.6 (spending less on a wrong guess\nproduces a wrong answer).\n\nThe key upgrades routing; it does not switch routing on. The classifier\nrouter is still opt-in (classifierRouter.enabled) and still needs a pool,\nbecause NeuroLink cannot invent the set of models you are willing to route\nbetween. What the key changes is which classifier runs inside a router you\nalready enabled.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Model routing","lvl3":""}},
 {"objectID":"40f783ceca290c039cce7f0199ab5758d7a0719b0a91c335c1c6e490a72b1412","title":"The model catalogue","url":"/docs/features/decide-inference-type#the-model-catalogue","content":"Enabling classifierRouter.catalog widens the routable pool beyond what you\ndeclared by hand: candidates are built from the 64-model registry (7\nproviders — see the model catalogue\nfor which), intersected with the credentials this host actually holds, and\nranked by a deterministic formula whenever no decision provider is available\nto choose among them.\n\n|                | Without catalog        | With catalog.enabled                |\n| -------------- | ------------------------ | ------------------------------------- |\n| Candidate pool | only the declared pool | declared pool plus registry matches |\n| Fallback pick  | first pool member        | tierScore()-ranked, tier-aware      |\n| Cap            | none needed              | maxModels, default 120              |\n\nSee the model catalogue for how\ncandidates are filtered, rendered, and ranked.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"The model catalogue","lvl3":""}},
 {"objectID":"1362463e9ab1e8a456cb34a4e5ac2ec8493f18ddc2a8d277282de858c0420131","title":"Per-request context budget","url":"/docs/features/decide-inference-type#per-request-context-budget","content":"A per-request compactionThreshold option lowers the point at which history\ngets compacted, below the 0.8-of-window default. The jev strategy can fill\nit in automatically from a four-level scope rubric (current-message through\neverything) — and the mapping is a one-directional invariant: it can only\never lower the 0.8 default, never raise it, because over-filling a window is\nan unrecoverable provider error.\n\nSee per-request context budget for the\nrubric, the invariant, and how the threshold scales the compaction target.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Per-request context budget","lvl3":""}},
@@ -3382,7 +3384,7 @@
 {"objectID":"b35e1ab2390a7b132037c293ecaf436f9ff9498591ccdd47b58ee0a6a95e7616","title":"RAG retrieval planning","url":"/docs/features/decide-inference-type#rag-retrieval-planning","content":"RAGPipelineConfig.decide lets each RAG query get its own topK/hybrid/\ngraph/rerank plan instead of one fixed configuration for every query. An\nexplicit per-call QueryOptions field always wins over the plan, and a\ncapability the pipeline wasn't configured with can never be switched on by\nit.\n\nSee per-query RAG retrieval planning\nfor the breadth rubric and why its confidence bar is deliberately lower than\ntool routing's or compaction's.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"RAG retrieval planning","lvl3":""}},
 {"objectID":"e399806055971566454170aeec7214ec9b5948c8e3a35731a9c64b4ee2be7c21","title":"One funnel for all five sites","url":"/docs/features/decide-inference-type#one-funnel-for-all-five-sites","content":"Every one of those calls is stamped with a site — routing, toolRouting,\ncontextRelevance, summaryGate or ragPlan (the DecisionSite union) — and\ngoes through one private path on NeuroLink, which is also where a bound\ntryDecide() lands. For the four sites that run inside a NeuroLink request —\nrouting, toolRouting, contextRelevance and summaryGate — that path\nforwards the outer generate() / stream() call's per-call credentials,\nabortSignal and requestId (the top-level option, else\ncontext.requestId) to the decision, so a request that carries\ncredentials: { laya: {...} } routes and compacts against that account\nrather than the instance's; for routing the per-call credentials also count\ntowards the auto strategy's activation, so a host that holds no decision key\nitself still routes a request that carries one. ragPlan runs inside a\nhost-wired RAGPipeline, outside any NeuroLink request, so it receives only\nwhat the host puts on the options it hands tryDecide (see \"Host-wired\nconsumers\" below). The path also adds decision.site and\ndecision.host_question_count to the model.decision span; emits\ndecision:before / decision:after on the instance emitter; and runs the\ndecisionHooks described next. Without hooks or listeners, the decision\npayload — state and questions — is unchanged; per-call credentials are still\nforwarded, so the full request on the wire (account, base URL) can differ from\nthe instance's.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"One funnel for all five sites","lvl3":""}},
 {"objectID":"a472e427cc594fbcc2ea01a87a8ebf99f49ebb52d71a7d5e6ff8baa6667707d1","title":"Riding along: decisionHooks","url":"/docs/features/decide-inference-type#riding-along-decisionhooks","content":"Latency on a decision model is flat in question count, so a host with its own\nquestions about the same request pays nothing to ask them in the round trip\nNeuroLink is already making. decisionHooks on the constructor does exactly\nthat — one request per site, never a second:\n\nWhat the funnel guarantees:\n\n| Concern                 | Behaviour                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |\n| ----------------------- |","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Riding along: decisionHooks","lvl3":""}},
-{"objectID":"fff45cefe577cc88ce4956ce2fbd51677da0f1d953c7c3330d96f4e26737da7e","title":"Reading the limits before you send: decisionLimits()","url":"/docs/features/decide-inference-type#reading-the-limits-before-you-send-decisionlimits","content":"neurolink.decisionLimits({ provider?, model?, credentials? }) returns the\nfigures the pre-flight check compares against, resolved for one model —\nprovider as decide() resolves it, model as model ?? the model decide() sends\n(the one registered for the provider: LAYA_MODEL / TYPESAFE_MODEL as read\nwhen providers were registered, else the default — a later change to the\nvariable moves neither), the per-model entry flattened over the base:\n\nmaxQuestions is present only for a provider that caps by question count\n(Laya); a reading without the key has no count cap, and a host should test\nfor the key rather than compare against Infinity. media is present only\nfor a provider that reads images or video (XOR, Perplexity, Cloudflare Clef,\nOpenAI Decisions); its absence is what the media refusal for TypeSafe and Laya is\nbased on.\n\nenforcedLocally says who refuses: NeuroLink itself, before any network call\n(Laya and XOR, whose servers would silently truncate or fail inside the\nengine), or the provider's server\n(TypeSafe and OpenAI Decisions, whose descriptor rows are marked advisory and are\nreported but never enforced here — a local refusal would turn a server ceiling\ninto a client one that drifts from it). estimateDecisionStateTokens is the very\nfunction the pre-flight refusal uses, so a state it sizes as fitting is sent\nand one it sizes as over is refused; pass the reading so non-ASCII text is\ncharged at the model's measured rate. A named provider is read whether or\nnot it is configured — its limits are facts about the model — while the\ndefault requires a configured provider, as tryDecide() does.\n\nLaya reads far less than Jev. About 768 tokens of state on\ntyped-decisions and multilingual, and 320 on english, auto or an\nunrecognised model name, against Jev's ~33,000. Laya's server does not refuse a\nlonger state; it answers from the first 1,024 tokens (512 on english).","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Reading the limits before you send: decisionLimits()","lvl3":""}},
+{"objectID":"fff45cefe577cc88ce4956ce2fbd51677da0f1d953c7c3330d96f4e26737da7e","title":"Reading the limits before you send: decisionLimits()","url":"/docs/features/decide-inference-type#reading-the-limits-before-you-send-decisionlimits","content":"neurolink.decisionLimits({ provider?, model?, credentials? }) returns the\nfigures the pre-flight check compares against, resolved for one model —\nprovider as decide() resolves it, model as model ?? the model decide() sends\n(the one registered for the provider: LAYA_MODEL / TYPESAFE_MODEL as read\nwhen providers were registered, else the default — a later change to the\nvariable moves neither), the per-model entry flattened over the base:\n\nmaxQuestions is present only for a provider that caps by question count\n(Laya); a reading without the key has no count cap, and a host should test\nfor the key rather than compare against Infinity. media is present only\nfor a provider that reads images or video (XOR, Perplexity, Cloudflare Clef,\nOpenAI Decisions); its absence is what the media refusal for TypeSafe, Laya, SubQ\nand ScaleDown is based on.\n\nenforcedLocally says who refuses: NeuroLink itself, before any network call\n(Laya and XOR, whose servers would silently truncate or fail inside the\nengine), or the provider's server\n(TypeSafe and OpenAI Decisions, whose descriptor rows are marked advisory and are\nreported but never enforced here — a local refusal would turn a server ceiling\ninto a client one that drifts from it). estimateDecisionStateTokens is the very\nfunction the pre-flight refusal uses, so a state it sizes as fitting is sent\nand one it sizes as over is refused; pass the reading so non-ASCII text is\ncharged at the model's measured rate. A named provider is read whether or\nnot it is configured — its limits are facts about the model — while the\ndefault requires a configured provider, as tryDecide() does.\n\nLaya reads far less than Jev. About 768 tokens of state on\ntyped-decisions and multilingual, and 320 on english, auto or an\nunrecognised model name, against Jev's ~33,000. Laya's server does not refuse a\nlonger state; it answers from the first 1,024 tokens (512 on english).","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Reading the limits before you send: decisionLimits()","lvl3":""}},
 {"objectID":"f6543a204326fbf7f27ac5c61c2f10b76f0541f3acb61eec93291921404e1a37","title":"What it is bad at","url":"/docs/features/decide-inference-type#what-it-is-bad-at","content":"\"Cannot hallucinate\" is a claim about output shape, not answer\ncorrectness: a decision model cannot return malformed JSON or an option you\ndid not offer, but it can still be wrong. TypeSafe reports ~68% accuracy on its\nown 711-case benchmark, against ~73% for a frontier model. It wins cost and\nlatency on every row and loses accuracy on every row — so it is right for\ndecisions that are gated and reversible, and wrong for final answers.\nIt reads literally. It answers the question you wrote, not the one you\n  meant. Split an ambiguous question into two and combine them in code.\nAsk about the act, not the subject. \"This task touches money\" scores high\n  on ordinary code that merely concerns money. The risk question in\n  classifyJev is worded to exclude writing and testing such code, precisely\n  because the naive phrasing escalated everything.\nIt is not a calculator. Counting, arithmetic and date comparison are\n  unreliable — dates are read as text, not ordered quantities. A score is for\n  thresholding and ranking, not for reading an exact magnitude off.\nIrrelevant state costs accuracy. Filter before sending.\nIt never explains itself. No rationale field exists, which rules it out\n  where a decision must be auditable.\nOption order can matter. Test with reordered criteria if a call is close.\nDon't invert criteria. A boolean whose true description means \"no\"\n  performs measurably worse.\nUntrusted input. Any text in state can steer a verdict: a user prompt,\n  a ticket, an earlier message, a retrieved document, tool output, a\n  transcript, a third-party server's description. A decision model can only\n  return an option you offered, but text written to push it can still pick the\n  option the author wants. The built-in consumers are advisory and fail-open,\n  and none of them feeds human-in-the-loop approval or a guardrail. Their reach\n  is bounded, not zero.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"What it is bad at","lvl3":""}},
 {"objectID":"f51228c123e6f86e1889f9990b2592debd24cc90c1584ad7d35d1f9a7aac8b82","title":"Adding another decision provider","url":"/docs/features/decide-inference-type#adding-another-decision-provider","content":"The decide inference type is provider-neutral by construction. Another\ndecision model needs:\nAn AIProviderName member and a <Name>Models enum\n   (src/lib/constants/enums.ts, outside the generated regions).\nA provider class extending SystemOneDecisionProvider\n   (src/lib/providers/systemOneDecision.ts), which owns the request loop,\n   retries, the auth circuit breaker and answer parsing, and supplies the\n   throwing getAISDKModel() / executeStream() stubs, the same shape the\n   embedding-only providers (voyage.ts, jina.ts) use. The class supplies\n   its endpoint, headers, body, error parsing and messages.\nA descriptor with inferenceKinds: [\"decide\"], no auto-select ranks, and\n   healthCheck: \"env-only\". That one field is what keeps a text-less model out\n   of every generation fallback chain; nothing else needs to know the provider\n   by name. A provider that reads images or video also declares\n   decisionLimits.media; without it, a request that carries media is refused\n   before any network call.\nA registration block, a credentials slice, a manifest, and the usual Tier-3\n   onboarding artifacts — pnpm run verify:provider-onboarding enumerates them.\n\nThe Tier-2 catalog JSON path cannot be used: its schema pins tier: 2, accepts\nonly an 8-flag text-generation capability vocabulary, and requires\ndefaultMaxOutputTokens and pricingPerMTok.output — none of which a model\nthat emits no text can honestly supply.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Adding another decision provider","lvl3":""}},
 {"objectID":"31d00aee561a454dc5082ad8002f9dbc0e33f15463cfee1636497043c8f9ee55","title":"Testing","url":"/docs/features/decide-inference-type#testing","content":"The suite drives dist/index.js only. Live tests skip without\nTYPESAFE_API_KEY; the degradation and discriminator tests run unconditionally,\nbecause \"behaves correctly with no key\" and \"a text-less provider is unreachable\nfrom generation\" are the contracts that matter most.","hierarchy":{"lvl0":"Features","lvl1":"The `decide` inference type","lvl2":"Testing","lvl3":""}},
@@ -3470,12 +3472,12 @@
 {"objectID":"acc039ba50dd533ebee67d761c7060a470341050b0e74b617d8a2ecd2ee58d3a","title":"File Size Limits","url":"/docs/features/file-processors#file-size-limits","content":"Default size limits prevent denial-of-service attacks:\n\n| Category     | Default Limit | Configurable     |\n| ------------ | ------------- | ---------------- |\n| Documents    | 50 MB         | Yes              |\n| Data files   | 10 MB         | Yes              |\n| Code files   | 5 MB          | Yes              |\n| Config files | 1 MB          | Yes              |\n| Images       | 10 MB         | No — fixed limit |\n\nThe image limit (SIZE_LIMITS_BYTES.IMAGE_MAX, 10 MB) is enforced: an oversized image buffer, file, or download throws a descriptive error before any base64 conversion, so a large image can no longer exhaust process memory. This applies to both entry points — images passed via generate({ files }) (routed through FileDetector → ImageProcessor) and via generate({ images }) (routed through the message builder). The internal image helpers that convert buffers/files/URLs accept an optional size override, but it is not exposed through generate() or any other public API — the 10 MB limit is fixed for SDK callers.","hierarchy":{"lvl0":"Features","lvl1":"File Processors Guide","lvl2":"File Size Limits","lvl3":""}},
 {"objectID":"65a3ceb133215b2453584713522686169451f444788d78d330508879d050037d","title":"Provider Compatibility","url":"/docs/features/file-processors#provider-compatibility","content":"All file processors work across NeuroLink's text- and multimodal-capable providers. The processed content is formatted as text that any such provider can understand:\n\n| Provider          | Documents | Data | Markup | Code | Config |\n| ----------------- | --------- | ---- | ------ | ---- | ------ |\n| OpenAI            | ✅        | ✅   | ✅     | ✅   | ✅     |\n| Anthropic         | ✅        | ✅   | ✅     | ✅   | ✅     |\n| Google AI Studio  | ✅        | ✅   | ✅     | ✅   | ✅     |\n| Google Vertex     | ✅        | ✅   | ✅     | ✅   | ✅     |\n| AWS Bedrock       | ✅        | ✅   | ✅     | ✅   | ✅     |\n| Azure OpenAI      | ✅        | ✅   | ✅     | ✅   | ✅     |\n| Mistral           | ✅        | ✅   | ✅     | ✅   | ✅     |\n| LiteLLM           | ✅        | ✅   | ✅     | ✅   | ✅     |\n| Ollama            | ✅        | ✅   | ✅     | ✅   | ✅     |\n| Hugging Face      | ✅        | ✅   | ✅     | ✅   | ✅     |\n| SageMaker         | ✅        | ✅   | ✅     | ✅   | ✅     |\n| OpenAI Compatible | ✅        | ✅   | ✅     | ✅   | ✅     |\n| OpenRouter        | ✅        | ✅   | ✅     | ✅   | ✅     |\n\nNote: For binary files like images and PDFs, provider-specific adapters handle the formatting. See PDF Support and Multimodal Chat.","hierarchy":{"lvl0":"Features","lvl1":"File Processors Guide","lvl2":"Provider Compatibility","lvl3":""}},
 {"objectID":"be7374d46193897235391ecd13b89a01971bdeb07cbb2205f0018d0d64f9d1bc","title":"Related Documentation","url":"/docs/features/file-processors#related-documentation","content":"Multimodal Chat - Image and media handling\nPDF Support - PDF-specific features\nCSV Support - CSV processing details\nCLI Commands - CLI file options\nSDK API Reference - Full API documentation","hierarchy":{"lvl0":"Features","lvl1":"File Processors Guide","lvl2":"Related Documentation","lvl3":""}},
-{"objectID":"b8b45370e9a5ddcc4d97fb6cfd40ec293136c5e986be4e0870e843a780c93877","title":"Guardrails AI Integration with Middleware","url":"/docs/features/guardrails-ai","content":"Guardrails AI Integration with Middleware\n\nThis document outlines the modern, simplified approach to integrating Guardrails AI with the NeuroLink platform using the new MiddlewareFactory. This enhances the safety, reliability, and security of your AI applications in a modular and maintainable way.\n\nOverview\n\nGuardrails AI is an open-source library that provides a framework for creating and managing guardrails for large language models (LLMs). By integrating Guardrails AI as middleware, you can enforce specific rules and policies on the inputs and outputs of your models, ensuring they adhere to your safety guidelines and quality standards.\n\nKey Benefits\nRisk Mitigation: Protect against common AI risks such as hallucinations, toxic language, and data leakage.\nQuality Assurance: Ensure that model outputs are accurate, relevant, and meet predefined quality criteria.\nCompliance: Enforce industry-specific regulations and compliance requirements.\nCustomization: Create custom guardrails tailored to specific use cases and business needs.\n\nMiddleware-based Guardrail Implementation\n\nWith the new MiddlewareFactory, integrating guardrails is easier than ever. The factory automatically handles the registration and application of the guardrails middleware when you use a relevant preset.\n\nUsing the security Preset\n\nThe easiest way to enable guardrails is to use the security preset when creating your MiddlewareFactory. This preset is specifically designed to enable the guardrails middleware with a default configuration.\n\nUsing the all Preset\n\nIf you want to use guardrails in combination with other built-in middleware like analytics, you can use the all preset.\n\nCustomizing Guardrails\n\nWhile presets provide a great starting point, you can also customize the behavior of the guardrails middleware by providing a custom configuration.\n\nThis new, streamlined approach provides a clean and scalable way to add safety and other enhancements to your AI models within the NeuroLink ecosystem.\n\nSee Also\n\nFor configuration examples, best practices, and troubleshooting, see the Guardrails Middleware Feature Guide.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails AI Integration with Middleware","lvl2":"","lvl3":""}},
+{"objectID":"b8b45370e9a5ddcc4d97fb6cfd40ec293136c5e986be4e0870e843a780c93877","title":"Guardrails AI Integration with Middleware","url":"/docs/features/guardrails-ai","content":"Guardrails AI Integration with Middleware\n\nThis document outlines the modern, simplified approach to integrating Guardrails AI with the NeuroLink platform using the new MiddlewareFactory. This enhances the safety, reliability, and security of your AI applications in a modular and maintainable way.\n\nOverview\n\nGuardrails AI is an open-source library that provides a framework for creating and managing guardrails for large language models (LLMs). By integrating Guardrails AI as middleware, you can enforce specific rules and policies on the inputs and outputs of your models, ensuring they adhere to your safety guidelines and quality standards.\n\nKey Benefits\nRisk Mitigation: Protect against common AI risks such as hallucinations, toxic language, and data leakage.\nQuality Assurance: Ensure that model outputs are accurate, relevant, and meet predefined quality criteria.\nCompliance: Enforce industry-specific regulations and compliance requirements.\nCustomization: Create custom guardrails tailored to specific use cases and business needs.\n\nMiddleware-based Guardrail Implementation\n\nWith the new MiddlewareFactory, integrating guardrails is easier than ever. The factory automatically handles the registration and application of the guardrails middleware when you use a relevant preset.\n\nUsing the security Preset\n\nThe easiest way to enable guardrails is to use the security preset when creating your MiddlewareFactory. This preset enables the guardrails middleware with an empty configuration: it filters nothing until you add badWords, modelFilter or precallEvaluation (see Customizing Guardrails). To apply a preset to every call without building a factory, use new NeuroLink({ middleware }), configureGlobalMiddleware() or NEUROLINK_MIDDLEWARE_PRESET.\n\nUsing the all Preset\n\nIf you want to use guardrails in combination with other built-in middleware like analytics, you can use the all preset.\n\nCustomizing Guardrails\n\nWhile presets provide a great starting point, you can also customize the behavior of the guardrails middleware by providing a custom configuration.\n\nThis new, streamlined approach provides a clean and scalable way to add safety and other enhancements to your AI models within the NeuroLink ecosystem.\n\nSee Also\n\nFor configuration examples, best practices, and troubleshooting, see the Guardrails Middleware Feature Guide.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails AI Integration with Middleware","lvl2":"","lvl3":""}},
 {"objectID":"2887e2921f080ab2b5f9cce892d0e5b231097704dada610addc7dba968ac2b2b","title":"Guardrails AI Integration with Middleware","url":"/docs/features/guardrails-ai#guardrails-ai-integration-with-middleware","content":"This document outlines the modern, simplified approach to integrating Guardrails AI with the NeuroLink platform using the new MiddlewareFactory. This enhances the safety, reliability, and security of your AI applications in a modular and maintainable way.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails AI Integration with Middleware","lvl2":"Guardrails AI Integration with Middleware","lvl3":""}},
 {"objectID":"656e3b1f93c26cedaeb4820d59bb33c0baf4c1d74f567b5e9903a4476b6df911","title":"Overview","url":"/docs/features/guardrails-ai#overview","content":"Guardrails AI is an open-source library that provides a framework for creating and managing guardrails for large language models (LLMs). By integrating Guardrails AI as middleware, you can enforce specific rules and policies on the inputs and outputs of your models, ensuring they adhere to your safety guidelines and quality standards.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails AI Integration with Middleware","lvl2":"Overview","lvl3":""}},
 {"objectID":"f2d2b26f599eb151126c2317daa9412449d66cc6b8b4d0f1327c8954b01c504a","title":"Key Benefits","url":"/docs/features/guardrails-ai#key-benefits","content":"Risk Mitigation: Protect against common AI risks such as hallucinations, toxic language, and data leakage.\nQuality Assurance: Ensure that model outputs are accurate, relevant, and meet predefined quality criteria.\nCompliance: Enforce industry-specific regulations and compliance requirements.\nCustomization: Create custom guardrails tailored to specific use cases and business needs.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails AI Integration with Middleware","lvl2":"Key Benefits","lvl3":""}},
 {"objectID":"26097f53ddbdab41d7341e08c5c758cbe1a3b95bbbbeae5dd96e9658aa979f2f","title":"Middleware-based Guardrail Implementation","url":"/docs/features/guardrails-ai#middleware-based-guardrail-implementation","content":"With the new MiddlewareFactory, integrating guardrails is easier than ever. The factory automatically handles the registration and application of the guardrails middleware when you use a relevant preset.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails AI Integration with Middleware","lvl2":"Middleware-based Guardrail Implementation","lvl3":""}},
-{"objectID":"5383993ee99d15ab0a7af3261df2e30969572d1867439cae6e2e66892657655a","title":"Using the security Preset","url":"/docs/features/guardrails-ai#using-the-security-preset","content":"The easiest way to enable guardrails is to use the security preset when creating your MiddlewareFactory. This preset is specifically designed to enable the guardrails middleware with a default configuration.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails AI Integration with Middleware","lvl2":"Using the security Preset","lvl3":""}},
+{"objectID":"5383993ee99d15ab0a7af3261df2e30969572d1867439cae6e2e66892657655a","title":"Using the security Preset","url":"/docs/features/guardrails-ai#using-the-security-preset","content":"The easiest way to enable guardrails is to use the security preset when creating your MiddlewareFactory. This preset enables the guardrails middleware with an empty configuration: it filters nothing until you add badWords, modelFilter or precallEvaluation (see Customizing Guardrails). To apply a preset to every call without building a factory, use new NeuroLink({ middleware }), configureGlobalMiddleware() or NEUROLINK_MIDDLEWARE_PRESET.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails AI Integration with Middleware","lvl2":"Using the security Preset","lvl3":""}},
 {"objectID":"422196a1c2e1508dcd31ea7fb0c1dba130bfbeb917925ceb9f6c80a762fcfd30","title":"Using the all Preset","url":"/docs/features/guardrails-ai#using-the-all-preset","content":"If you want to use guardrails in combination with other built-in middleware like analytics, you can use the all preset.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails AI Integration with Middleware","lvl2":"Using the all Preset","lvl3":""}},
 {"objectID":"4b7a1fe97ea0f23a386d90548d3214e8e24a6f36ffcff4df7d5859a78e6f9271","title":"Customizing Guardrails","url":"/docs/features/guardrails-ai#customizing-guardrails","content":"While presets provide a great starting point, you can also customize the behavior of the guardrails middleware by providing a custom configuration.\n\nThis new, streamlined approach provides a clean and scalable way to add safety and other enhancements to your AI models within the NeuroLink ecosystem.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails AI Integration with Middleware","lvl2":"Customizing Guardrails","lvl3":""}},
 {"objectID":"1c07c45bcdcf7be8441182379a01af46db12b22af1ce3115ca29218979bd43bb","title":"See Also","url":"/docs/features/guardrails-ai#see-also","content":"For configuration examples, best practices, and troubleshooting, see the Guardrails Middleware Feature Guide.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails AI Integration with Middleware","lvl2":"See Also","lvl3":""}},
@@ -3506,32 +3508,34 @@
 {"objectID":"062ba86ed5fa937d989ef36219f6bc4d2673654750f6c4fbfee623527282bc03","title":"From Previous Implementations","url":"/docs/features/guardrails-implementation#from-previous-implementations","content":"If upgrading from older guardrail implementations:\nUpdate configuration format to new interfaces\nReplace deprecated methods with new utility functions\nTest evaluation thresholds and adjust as needed\nUpdate error handling to use new patterns","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Implementation Guide","lvl2":"From Previous Implementations","lvl3":""}},
 {"objectID":"1d4cfe18f157b91540119afedbf464d626c2c320e56551b69d5b3f853541e3c4","title":"Breaking Changes","url":"/docs/features/guardrails-implementation#breaking-changes","content":"Configuration structure has been updated for better organization\nSome utility function signatures have changed\nError handling patterns have been improved","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Implementation Guide","lvl2":"Breaking Changes","lvl3":""}},
 {"objectID":"214fe14d9f320f468b5ed10d5b5cf315a7aa2a0fcb1a32598f42a622570e177d","title":"Conclusion","url":"/docs/features/guardrails-implementation#conclusion","content":"The NeuroLink guardrails implementation provides comprehensive content safety and filtering capabilities with:\n✅ AI-powered pre-call evaluation\n✅ Flexible content filtering options\n✅ Configurable response actions\n✅ Visual proof and demonstrations\n✅ High performance and scalability\n✅ Comprehensive error handling\n✅ TypeScript support throughout\n\nFor additional support or questions, refer to the main NeuroLink documentation or create an issue in the repository.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Implementation Guide","lvl2":"Conclusion","lvl3":""}},
-{"objectID":"540c47bd7c6e059a9673d6acb056728c913d14a87db9e96c85e9ad17847af52d","title":"Guardrails Middleware","url":"/docs/features/guardrails","content":"Guardrails Middleware\n\nSince: v7.42.0 | Status: Stable | Availability: SDK (CLI + SDK)\n\nOverview\n\nWhat it does: Guardrails middleware provides real-time content filtering and policy enforcement for AI model outputs, blocking profanity, PII, unsafe content, and custom-defined terms.\n\nWhy use it: Protect your application from generating harmful, inappropriate, or non-compliant content. Ensures AI responses meet safety standards and regulatory requirements.\n\nCommon use cases:\nContent moderation for user-facing applications\nPII (Personally Identifiable Information) redaction\nProfanity filtering for family-friendly apps\nCompliance with industry regulations (COPPA, GDPR, etc.)\nBrand safety and reputation management\n\nQuick Start\n\nGuardrails work out of the box with the security preset. No custom configuration required for basic content filtering.\n\nSDK Example with Security Preset\nEnables guardrails middleware with default configuration\nPass middleware on every call that should be filtered; the NeuroLink constructor does not take this option\nContent is already filtered - safe to display to users\n\nCustom Guardrails Configuration\nMaster switch for guardrails middleware\nEnable keyword-based filtering (fast, regex-based)\nCustom terms to filter/redact from outputs\nEnable AI-powered content safety check (slower, more accurate)\nUse fast, cheap model for safety evaluation\n\nCLI Usage\n\nConfiguration\n\n| Option                    | Type       | Default | Required | Description                          |\n| ------------------------- | ---------- | ------- | -------- | ------------------------------------ |\n| enabled                 | boolean  | true  | No       | Enable/disable guardrails middleware |\n| badWords.enabled        | boolean  | false | No       | Enable keyword-based filtering       |\n| badWords.list           | string[] | []    | No       | List of terms to filter/redact       |\n| badWords.regexPatterns  | string[] | []    | No       | Regular expressions to redact        |\n| modelFilter.enabled     | boolean  | false | No       | Enable AI-based content safety check |\n| modelFilter.filterModel | string   | -       | No       | Model to use for safety evaluation   |\n\nEnvironment Variables\n\nConfig File\n\nHow It Works\n\nFiltering Pipeline\nUser prompt → Sent to AI model\nAI generates response → Initial content created\nGuardrails middleware intercepts:\nBad word filtering: Regex-based term replacement\nModel-based filtering: AI evaluates content safety\nFiltered response → Delivered to user\n\nBad Word Filtering\n\nSimple regex-based replacement:\nCase-insensitive matching\nReplaces each match with a fixed string, [REDACTED] by default (not length-preserving asterisks) — override it with badWords.replacementText\nWorks in both generate and stream modes\n\nRegex Patterns\n\nbadWords.regexPatterns takes regular expressions, each applied with the gi flags. When it is set it takes priority over list:\n\nPatterns run synchronously against the whole response, so one that backtracks catastrophically would stall every request in the process. The cost comes from input that can be split between the parts of a pattern in many ways: a variable-length part (+, *, ?, {n,m}) that can match the same text as whatever follows it. A pattern is refused when it:\nis longer than 1000 characters;\nrepeats a group in which such a part can match the same text as what follows it, including the next iteration of the group: (a+)+, (\\w+\\s*)*, (.*a){10}, (a?a?)+, (a{1,30}){1,30};\nrepeats a group whose alternatives may overlap, such as (a|aa)+ or (\\w|\\d)+;\nchains too many overlapping variable-length parts: more than 8 optional parts in a row (a?a?a?a?a?a?a?a?a?), or more than 4 open-ended ones (.*a.*b.*c.*d.*e.*f);\nuses a repetition bound above 500, such as a{1,1000}.\n\nA repeated group is fine when each variable-length part is followed by something it cannot match. These all run: (\\d{1,3}\\.){3}, ([a-z0-9-]+\\.)+[a-z]{2,}, (?:\\d{4}[- ]?)+, \\b(?:\\d[ -]*?){13,16}\\b, (?:\\w+\\s+)+.\n\nA refused pattern is skipped and logged as an error (Refusing regex pattern ...), the same way an invalid regex is. It is never thrown at the caller, and the other patterns in the list still run. Fail-open is the trade: a skipped pattern filters nothing, so check the logs after you change a pattern. The same limits apply to precallEvaluation.sanitizationPatterns (Refusing sanitization pattern ...).\n\nThe check reads the pattern; it does not run it, so it is conservative and not a proof:\nIt refuses some safe patterns, such as a small fixed count of an ambiguous part ((a{1,3}){3}). Flatten it (a{1,9}) or split it.\nIt compares character classes on a fixed sample of characters (Latin text plus a few other scripts), so a class that only overlaps outside the sample can look disjoint. A back-reference is assumed to overlap with everything.\nExponential shapes built from optional parts are only partly covered: loops and long chains are refused, but the check is a model of the engine, not the engine. Open-ende","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"","lvl3":""}},
-{"objectID":"ad4c8c0e011c259d8f444afec90c94a3c939d9f14e65491e8044a8417430b54c","title":"Guardrails Middleware","url":"/docs/features/guardrails#guardrails-middleware","content":"Since: v7.42.0 | Status: Stable | Availability: SDK (CLI + SDK)","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Guardrails Middleware","lvl3":""}},
-{"objectID":"1ecba9f0cdb3586cf8ad53bba31fb80cabab7aed2aba88050b4f98078dd2bec9","title":"Overview","url":"/docs/features/guardrails#overview","content":"What it does: Guardrails middleware provides real-time content filtering and policy enforcement for AI model outputs, blocking profanity, PII, unsafe content, and custom-defined terms.\n\nWhy use it: Protect your application from generating harmful, inappropriate, or non-compliant content. Ensures AI responses meet safety standards and regulatory requirements.\n\nCommon use cases:\nContent moderation for user-facing applications\nPII (Personally Identifiable Information) redaction\nProfanity filtering for family-friendly apps\nCompliance with industry regulations (COPPA, GDPR, etc.)\nBrand safety and reputation management","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Overview","lvl3":""}},
-{"objectID":"064824951ceeb2f827852673ce43e1d5b0a4776cda59dec766413ea3ea1a41ae","title":"Quick Start","url":"/docs/features/guardrails#quick-start","content":"Guardrails work out of the box with the security preset. No custom configuration required for basic content filtering.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Quick Start","lvl3":""}},
-{"objectID":"95afe6ceda57115c0508134d07f22b3d8ca5fa5f064adacafde697b1e97a1f49","title":"SDK Example with Security Preset","url":"/docs/features/guardrails#sdk-example-with-security-preset","content":"Enables guardrails middleware with default configuration\nPass middleware on every call that should be filtered; the NeuroLink constructor does not take this option\nContent is already filtered - safe to display to users","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"SDK Example with Security Preset","lvl3":""}},
-{"objectID":"b4fbe7f14d6486c89e71170dde75c0aac0a463e7146d3505542a389e27177987","title":"Custom Guardrails Configuration","url":"/docs/features/guardrails#custom-guardrails-configuration","content":"Master switch for guardrails middleware\nEnable keyword-based filtering (fast, regex-based)\nCustom terms to filter/redact from outputs\nEnable AI-powered content safety check (slower, more accurate)\nUse fast, cheap model for safety evaluation","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Custom Guardrails Configuration","lvl3":""}},
-{"objectID":"c8521792297f968b1ed06d4d84924590f9453bb688d0511d82b24e8cb807d1d4","title":"CLI Usage","url":"/docs/features/guardrails#cli-usage","content":"`bash","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"CLI Usage","lvl3":""}},
-{"objectID":"db6d627e8bf12acc503a42180b82237d38b98a34bf4b5c6427110388ba783598","title":"Enable guardrails via environment variable","url":"/docs/features/guardrails#enable-guardrails-via-environment-variable","content":"npx @juspay/neurolink generate \"Write a product description\" --enable-analytics","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Enable guardrails via environment variable","lvl3":""}},
-{"objectID":"0968c358bc16ec092d8eafc1875703e50dd25b900e6a64d8f9a1356ac3e143f8","title":"Guardrails are automatically applied to all generations","url":"/docs/features/guardrails#guardrails-are-automatically-applied-to-all-generations","content":"`","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Guardrails are automatically applied to all generations","lvl3":""}},
-{"objectID":"5b562e23dda525ed54a3fbcd8d86fd0b6b8aba8bc84bb34bc52b2cd028cf0178","title":"Configuration","url":"/docs/features/guardrails#configuration","content":"| Option                    | Type       | Default | Required | Description                          |\n| ------------------------- | ---------- | ------- | -------- | ------------------------------------ |\n| enabled                 | boolean  | true  | No       | Enable/disable guardrails middleware |\n| badWords.enabled        | boolean  | false | No       | Enable keyword-based filtering       |\n| badWords.list           | string[] | []    | No       | List of terms to filter/redact       |\n| badWords.regexPatterns  | string[] | []    | No       | Regular expressions to redact        |\n| modelFilter.enabled     | boolean  | false | No       | Enable AI-based content safety check |\n| modelFilter.filterModel | string   | -       | No       | Model to use for safety evaluation   |","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Configuration","lvl3":""}},
-{"objectID":"2406e4412d40d6ab610c334f551ae73529304f50a559b6fdf180d616d7536174","title":"Environment Variables","url":"/docs/features/guardrails#environment-variables","content":"`bash","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Environment Variables","lvl3":""}},
-{"objectID":"946b41af2f888ce1a40149ea8123d2c4584292d7a22fe88549f0b7c23671f951","title":"Or enable all middleware (includes guardrails + analytics)","url":"/docs/features/guardrails#or-enable-all-middleware-includes-guardrails-analytics","content":"`","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Or enable all middleware (includes guardrails + analytics)","lvl3":""}},
-{"objectID":"70cbcd4ca56befe769c6ddb6128a5ab280c75a48edc8e9b4141dfce607bc2c1e","title":"Filtering Pipeline","url":"/docs/features/guardrails#filtering-pipeline","content":"User prompt → Sent to AI model\nAI generates response → Initial content created\nGuardrails middleware intercepts:\nBad word filtering: Regex-based term replacement\nModel-based filtering: AI evaluates content safety\nFiltered response → Delivered to user","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Filtering Pipeline","lvl3":""}},
-{"objectID":"3af7a22581a93723a5045560d7c0026a6fc3f1bca903dd47c28e0dfa5707e350","title":"Bad Word Filtering","url":"/docs/features/guardrails#bad-word-filtering","content":"Simple regex-based replacement:\nCase-insensitive matching\nReplaces each match with a fixed string, [REDACTED] by default (not length-preserving asterisks) — override it with badWords.replacementText\nWorks in both generate and stream modes","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Bad Word Filtering","lvl3":""}},
+{"objectID":"540c47bd7c6e059a9673d6acb056728c913d14a87db9e96c85e9ad17847af52d","title":"Guardrails Middleware","url":"/docs/features/guardrails","content":"Guardrails Middleware\n\nSince: v7.29.0 | Status: Stable | Availability: SDK; the CLI reads NEUROLINK_MIDDLEWARE_PRESET\n\nOverview\n\nWhat it does: Guardrails is a built-in language-model middleware (id guardrails) that checks what a model returns, and optionally what a user sends. It has three mechanisms. Each is off until you configure it:\nBad-word filtering replaces matches of terms or regular expressions you list with a fixed string. Works in generate() and stream().\nModel-based filtering asks a second model whether the reply is safe and replaces the whole reply if the answer is unsafe. generate() only.\nPre-call evaluation asks a model to judge the user's input before the request is sent, and answers a blocked request with a fixed refusal. Works in generate() and stream().\n\nWhat it does not do: Guardrails ships no word list, no profanity list and no PII patterns. With nothing configured it changes nothing, including under preset: \"security\". It is a text filter, not a compliance control: a term that is not in your list passes through, and a model-based check can miss unsafe content or be unavailable (see Model-Based Filtering). Built-in PII detection is a separate piiDetection option on generate() and stream(), not part of guardrails.\n\nCommon use cases:\nRedacting a known set of terms (brand names, internal codenames, banned words) from model output\nA second-opinion safety check on replies from generate()\nRefusing requests that an evaluator model judges unsafe before they reach the main model\n\nQuick Start\n\nRedact terms on one call\n\nBoth middlewareConfig.guardrails.enabled and badWords.enabled must be true, and list or regexPatterns must be non-empty. Missing any one of them filters nothing.\n\nSet it once for an instance\n\nThe preset: \"security\" line is what switches guardrails on here; the word list is what makes it filter.\n\nSet it once for the process\n\nTurn a preset on from the environment\n\npreset: \"security\" only enables the guardrails middleware. It adds no badWords and no modelFilter, so until one of the layers below supplies them, a reply passes through unchanged. That is true of NEUROLINK_MIDDLEWARE_PRESET=security too. There is deliberately no default word list: a generic list gives false assurance and false positives.\n\nDefaults and precedence\n\nA call resolves its middleware from four layers. A higher layer overrides a lower one.\n\n| Order       | Layer       | How you set it                       | Scope                                                                                                            |\n| ----------- | ----------- | ------------------------------------ | ---------------------------------------------------------------------------------------------------------------- |\n| 1 (lowest)  | Environment | NEUROLINK_MIDDLEWARE_PRESET        | Read when each NeuroLink instance is constructed. A preset name only: default, all or security           |\n| 2           | Global      | configureGlobalMiddleware(options) | The whole process. Read on every call, so existing instances pick it up. A later call replaces the earlier one   |\n| 3           | Instance    | new NeuroLink({ middleware })      | That instance. Snapshotted at construction: later edits to your object's word lists or strings are not picked up |\n| 4 (highest) | Per call    | generate({ middleware }) and so on | That call                                                                                                        |\n\nThe defaults apply to generate(), stream() and generateText(). With no default set, a call behaves exactly as it did before: only its own middleware counts.\n\nThe classifier router's model strategy and the probe call in checkCredentials() run through generate() on your own instance, so all of that instance's defaults apply to them. The SDK also creates separate NeuroLink instances for the file summariser, the conversation-memory condenser, the Redis conversation title generator and the image-generation service. Those see the environment layer, which every instance reads, and the global layer, which is read on every call, but not your instance's middleware option. A global badWords list therefore redacts file summaries and titles, and a global precallEvaluation sends whole file contents to the evaluator model.\n\nAn unrecognised NEUROLINK_MIDDLEWARE_PRESET is ignored. A message naming the value is written to stderr, once per distinct value, whether or not debugging is on. A preset in a constructor or global default that is not default, all or security throws a TypeError at new NeuroLink() or configureGlobalMiddleware(), so a typo cannot silently replace a valid preset from a lower layer and leave guardrails off.\n\nHow layers combine\n\n| Option                                                          | Rule                                                                                                                                |\n| --------------------------------------------------------------- | ----------------","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"","lvl3":""}},
+{"objectID":"ad4c8c0e011c259d8f444afec90c94a3c939d9f14e65491e8044a8417430b54c","title":"Guardrails Middleware","url":"/docs/features/guardrails#guardrails-middleware","content":"Since: v7.29.0 | Status: Stable | Availability: SDK; the CLI reads NEUROLINK_MIDDLEWARE_PRESET","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Guardrails Middleware","lvl3":""}},
+{"objectID":"1ecba9f0cdb3586cf8ad53bba31fb80cabab7aed2aba88050b4f98078dd2bec9","title":"Overview","url":"/docs/features/guardrails#overview","content":"What it does: Guardrails is a built-in language-model middleware (id guardrails) that checks what a model returns, and optionally what a user sends. It has three mechanisms. Each is off until you configure it:\nBad-word filtering replaces matches of terms or regular expressions you list with a fixed string. Works in generate() and stream().\nModel-based filtering asks a second model whether the reply is safe and replaces the whole reply if the answer is unsafe. generate() only.\nPre-call evaluation asks a model to judge the user's input before the request is sent, and answers a blocked request with a fixed refusal. Works in generate() and stream().\n\nWhat it does not do: Guardrails ships no word list, no profanity list and no PII patterns. With nothing configured it changes nothing, including under preset: \"security\". It is a text filter, not a compliance control: a term that is not in your list passes through, and a model-based check can miss unsafe content or be unavailable (see Model-Based Filtering). Built-in PII detection is a separate piiDetection option on generate() and stream(), not part of guardrails.\n\nCommon use cases:\nRedacting a known set of terms (brand names, internal codenames, banned words) from model output\nA second-opinion safety check on replies from generate()\nRefusing requests that an evaluator model judges unsafe before they reach the main model","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Overview","lvl3":""}},
+{"objectID":"6215004e045d98a4935173857d1228a9c7699746998c801501ac8f9b9de0dfef","title":"Redact terms on one call","url":"/docs/features/guardrails#redact-terms-on-one-call","content":"Both middlewareConfig.guardrails.enabled and badWords.enabled must be true, and list or regexPatterns must be non-empty. Missing any one of them filters nothing.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Redact terms on one call","lvl3":""}},
+{"objectID":"dd7d7f1a5d3cff8d00a5d56133e0a8345ff1eb7873a6a7b12239af036a7f828e","title":"Set it once for an instance","url":"/docs/features/guardrails#set-it-once-for-an-instance","content":"The preset: \"security\" line is what switches guardrails on here; the word list is what makes it filter.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Set it once for an instance","lvl3":""}},
+{"objectID":"3cd4c72a4af5c6c901a4261fd6506c27591d7b3d3f7b5ee6875af27182613eac","title":"Turn a preset on from the environment","url":"/docs/features/guardrails#turn-a-preset-on-from-the-environment","content":"preset: \"security\" only enables the guardrails middleware. It adds no badWords and no modelFilter, so until one of the layers below supplies them, a reply passes through unchanged. That is true of NEUROLINK_MIDDLEWARE_PRESET=security too. There is deliberately no default word list: a generic list gives false assurance and false positives.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Turn a preset on from the environment","lvl3":""}},
+{"objectID":"b704c7417bcc49bb43075e7000d652885dd9d1ffb780f82dd398fef4861c6584","title":"Defaults and precedence","url":"/docs/features/guardrails#defaults-and-precedence","content":"A call resolves its middleware from four layers. A higher layer overrides a lower one.\n\n| Order       | Layer       | How you set it                       | Scope                                                                                                            |\n| ----------- | ----------- | ------------------------------------ | ---------------------------------------------------------------------------------------------------------------- |\n| 1 (lowest)  | Environment | NEUROLINK_MIDDLEWARE_PRESET        | Read when each NeuroLink instance is constructed. A preset name only: default, all or security           |\n| 2           | Global      | configureGlobalMiddleware(options) | The whole process. Read on every call, so existing instances pick it up. A later call replaces the earlier one   |\n| 3           | Instance    | new NeuroLink({ middleware })      | That instance. Snapshotted at construction: later edits to your object's word lists or strings are not picked up |\n| 4 (highest) | Per call    | generate({ middleware }) and so on | That call                                                                                                        |\n\nThe defaults apply to generate(), stream() and generateText(). With no default set, a call behaves exactly as it did before: only its own middleware counts.\n\nThe classifier router's model strategy and the probe call in checkCredentials() run through generate() on your own instance, so all of that instance's defaults apply to them. The SDK also creates separate NeuroLink instances for the file summariser, the conversation-memory condenser, the Redis conversation title generator and the image-generation service. Those see the environment layer, which every instance reads, and the global layer, which is read on every call, but not your instance's middleware option. A global badWords list therefore redacts file summaries and titles, and a global precallEvaluation sends whole file contents to the evaluator model.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Defaults and precedence","lvl3":""}},
+{"objectID":"1166b98bef204d316c02e3d1088946a13fe15472dbe6e89fbbe3d069fc80cd9b","title":"How layers combine","url":"/docs/features/guardrails#how-layers-combine","content":"| Option                                                          | Rule                                                                                                                                |\n| --------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------- |\n| preset                                                        | The highest layer that sets one wins                                                                                                |\n| middlewareConfig                                              | Merged per middleware id. Within an id, enabled and conditions come from the highest layer that sets them                       |\n| middlewareConfig.<id>.config                                  | Merged key by key. A key set by a higher layer replaces the lower layer's value for that key as a whole: badWords is not combined |\n| middleware, enabledMiddleware, disabledMiddleware (lists) | The highest layer that sets the list replaces it. Lists are never concatenated. An explicit [] clears an inherited list           |\n| global                                                        | Merged key by key                                                                                                                   |\n\nThe factory applies disabledMiddleware after enabledMiddleware, so a lower layer's disabled list could out-vote a call that asked for the opposite. The merge prevents that: an id that a higher layer switches on, by enabledMiddleware or by middlewareConfig.<id>.enabled: true, is removed from a disabled list it would inherit, and the reverse.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"How layers combine","lvl3":""}},
+{"objectID":"5b562e23dda525ed54a3fbcd8d86fd0b6b8aba8bc84bb34bc52b2cd028cf0178","title":"Configuration","url":"/docs/features/guardrails#configuration","content":"These go under middlewareConfig.guardrails.config, except the first row.\n\n| Option                                | Type              | Default                             | Description                                                                                                         |\n| ------------------------------------- | ----------------- | ----------------------------------- | ------------------------------------------------------------------------------------------------------------------- |\n| middlewareConfig.guardrails.enabled | boolean         | false, unless a preset enables it | Master switch. preset: \"security\" and preset: \"all\" set it to true                                            |\n| badWords.enabled                    | boolean         | false                             | Turn on term filtering                                                                                              |\n| badWords.list                       | string[]        | none                                | Terms to redact. Case-insensitive, matched as substrings (no word boundaries), so ass also matches inside class |\n| badWords.regexPatterns              | string[]        | none                                | Case-insensitive regular expressions. When both this and list are set, list is ignored and a warning is logged  |\n| badWords.replacementText            | string          | [REDACTED]                        | What each match is replaced with. The replacement is a fixed string, not length-preserving                          |\n| modelFilter.enabled                 | boolean         | false                             | Turn on the model-based check (generate() only)                                                                   |\n| modelFilter.filterModel             | string or model | none                                | \"provider:model\", a bare model id, or a model handle.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Configuration","lvl3":""}},
+{"objectID":"2406e4412d40d6ab610c334f551ae73529304f50a559b6fdf180d616d7536174","title":"Environment variables","url":"/docs/features/guardrails#environment-variables","content":"| Variable                      | Values                       | Effect                                                                                                                         |\n| ----------------------------- | ---------------------------- | ------------------------------------------------------------------------------------------------------------------------------ |\n| NEUROLINK_MIDDLEWARE_PRESET | default, all, security | Lowest-precedence default preset for every NeuroLink instance. Blank is ignored; unknown is ignored with a message on stderr |\n\n| Preset     | Turns on              | Notes                                                                         |\n| ---------- | --------------------- | ----------------------------------------------------------------------------- |\n| default  | analytics             |                                                                               |\n| all      | analytics, guardrails | Not everything: auto-evaluation and lifecycle are opt-in per call             |\n| security | guardrails            | Filters nothing until badWords, modelFilter or precallEvaluation is set |\n\nThere is no config-file setting for middleware. Use the constructor option, configureGlobalMiddleware() or the environment variable.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Environment variables","lvl3":""}},
+{"objectID":"71eea9da1fd111fc54519592dbdef82d137cfc893048f58a83e18650fb783ffb","title":"CLI","url":"/docs/features/guardrails#cli","content":"The CLI builds a NeuroLink instance, so NEUROLINK_MIDDLEWARE_PRESET applies to CLI runs too:\n\nThe CLI has no option for a word list or a model filter, so this enables guardrails and filters nothing. Use the SDK to choose what is filtered.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"CLI","lvl3":""}},
+{"objectID":"70cbcd4ca56befe769c6ddb6128a5ab280c75a48edc8e9b4141dfce607bc2c1e","title":"Filtering Pipeline","url":"/docs/features/guardrails#filtering-pipeline","content":"Your request is sent to the model. With precallEvaluation on, an evaluator model sees the input first and an unsafe request is answered with Request contains inappropriate content and has been blocked. without reaching the main model.\nThe model replies.\nThe guardrails middleware filters the reply:\nBad words: the configured terms or patterns are replaced.\nModel filter (generate() only): a second model classifies the reply.\nThe filtered reply is delivered.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Filtering Pipeline","lvl3":""}},
+{"objectID":"3af7a22581a93723a5045560d7c0026a6fc3f1bca903dd47c28e0dfa5707e350","title":"Bad Word Filtering","url":"/docs/features/guardrails#bad-word-filtering","content":"Case-insensitive\nReplaces each match with a fixed string, [REDACTED] by default; override it with badWords.replacementText\nA term matches anywhere inside a word. Use regexPatterns with \\b when you need word boundaries\nWorks in both generate() and stream()","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Bad Word Filtering","lvl3":""}},
 {"objectID":"4be9b0134609c0a5ccba707be47bfc7c5708940e325c69550c196de6060f48a0","title":"Regex Patterns","url":"/docs/features/guardrails#regex-patterns","content":"badWords.regexPatterns takes regular expressions, each applied with the gi flags. When it is set it takes priority over list:\n\nPatterns run synchronously against the whole response, so one that backtracks catastrophically would stall every request in the process. The cost comes from input that can be split between the parts of a pattern in many ways: a variable-length part (+, *, ?, {n,m}) that can match the same text as whatever follows it. A pattern is refused when it:\nis longer than 1000 characters;\nrepeats a group in which such a part can match the same text as what follows it, including the next iteration of the group: (a+)+, (\\w+\\s*)*, (.*a){10}, (a?a?)+, (a{1,30}){1,30};\nrepeats a group whose alternatives may overlap, such as (a|aa)+ or (\\w|\\d)+;\nchains too many overlapping variable-length parts: more than 8 optional parts in a row (a?a?a?a?a?a?a?a?a?), or more than 4 open-ended ones (.*a.*b.*c.*d.*e.*f);\nuses a repetition bound above 500, such as a{1,1000}.\n\nA repeated group is fine when each variable-length part is followed by something it cannot match. These all run: (\\d{1,3}\\.){3}, ([a-z0-9-]+\\.)+[a-z]{2,}, (?:\\d{4}[- ]?)+, \\b(?:\\d[ -]*?){13,16}\\b, (?:\\w+\\s+)+.\n\nA refused pattern is skipped and logged as an error (Refusing regex pattern ...), the same way an invalid regex is. It is never thrown at the caller, and the other patterns in the list still run. Fail-open is the trade: a skipped pattern filters nothing, so check the logs after you change a pattern. The same limits apply to precallEvaluation.sanitizationPatterns (Refusing sanitization pattern ...).\n\nThe check reads the pattern; it does not run it, so it is conservative and not a proof:\nIt refuses some safe patterns, such as a small fixed count of an ambiguous part ((a{1,3}){3}). Flatten it (a{1,9}) or split it.\nIt compares character classes on a fixed sample of characters (Latin text plus a few other scripts), so a class that only overlaps outside the sample can look disjoint.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Regex Patterns","lvl3":""}},
-{"objectID":"5c5b50314493905262b204dbb73bacc7c6ab8a60fa3124469df9e6c1706e47a5","title":"Model-Based Filtering","url":"/docs/features/guardrails#model-based-filtering","content":"While guardrails filter common PII patterns, always review critical outputs manually. False negatives can occur with obfuscated data or uncommon PII formats. For high-stakes compliance, combine with dedicated PII detection services.\n\nAI-powered safety check:\nUses separate, lightweight model (e.g., gpt-4o-mini)\nBinary safe/unsafe classification\nFull redaction on unsafe detection","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Model-Based Filtering","lvl3":""}},
-{"objectID":"dbfe7ff9f27039591c8f98fc3b03c705f22f7b36b689e5f0de95241ae3fb71ee","title":"Middleware Configuration","url":"/docs/features/guardrails#middleware-configuration","content":"preset: \"security\" → Enables guardrails with defaults\npreset: \"all\" → Enables guardrails + all other middleware\nmiddlewareConfig.guardrails → Custom guardrails configuration\n\nSee guardrails-ai-integration.md for complete integration guide.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Middleware Configuration","lvl3":""}},
-{"objectID":"85995a92035eb64e3e92878f16c818aa6b02432dfcdf8a2c1b7d0ec93efb0ceb","title":"Problem: Guardrails not filtering content","url":"/docs/features/guardrails#problem-guardrails-not-filtering-content","content":"Cause: Middleware not enabled or preset not configured\nSolution:","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Problem: Guardrails not filtering content","lvl3":""}},
-{"objectID":"69697f2c3191cc2ac3f2d5a0a1e9295a7c6df79603d7a54068a65420afd69858","title":"Problem: Too many false positives (legitimate content filtered)","url":"/docs/features/guardrails#problem-too-many-false-positives-legitimate-content-filtered","content":"Cause: Overly aggressive bad word list\nSolution:","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Problem: Too many false positives (legitimate content filtered)","lvl3":""}},
+{"objectID":"03371e64b1c11c9930bf187a72312fff1e1954628c7da4981da6c3507dfa150b","title":"Streaming","url":"/docs/features/guardrails#streaming","content":"With badWords.enabled: true, stream() holds each run of text and filters it as one string, so a term split across chunks is still caught. The run is released when a non-text part arrives or the stream ends. Filtering therefore costs incremental delivery: you may receive a long reply in one piece at the end rather than as it is generated. With bad-word filtering off, chunks pass through unbuffered.\n\nneurolink.streamText(prompt, options?) is the older string-only form: it takes the prompt as its first argument and yields plain strings, not objects with a content field.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Streaming","lvl3":""}},
+{"objectID":"5c5b50314493905262b204dbb73bacc7c6ab8a60fa3124469df9e6c1706e47a5","title":"Model-Based Filtering","url":"/docs/features/guardrails#model-based-filtering","content":"After the main model replies, the text is sent to the filter model with the prompt Is the following text safe? Respond with only \"safe\" or \"unsafe\". If the answer, trimmed and lower-cased, is exactly unsafe, each text part of the reply is replaced by <REDACTED BY AI GUARDRAIL>. Any other answer leaves the reply alone.\ngenerate() only. stream() does not run the model filter.\nfilterModel as a string is \"provider:model\". A bare model id such as gpt-4o-mini is created on the default provider (NEUROLINK_PROVIDER, then AI_PROVIDER, then vertex), so name the provider unless that is what you want.\nIt fails open. If the filter model cannot be created or the call fails, an error is logged and the reply is returned unfiltered.\nThe reply text leaves for another model. The filter model's provider sees what the main model said.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Model-Based Filtering","lvl3":""}},
+{"objectID":"7ff25d1d098f3d16de4560efbe63e65954883804b47cb56ebc73307b7582f6e3","title":"Pre-Call Evaluation","url":"/docs/features/guardrails#pre-call-evaluation","content":"With precallEvaluation.enabled: true, the user's input is sent to an evaluator model (google-ai with gemini-2.5-flash unless you set provider and evaluationModel). An input it judges unsafe is blocked by default; if the evaluator errors, the request is allowed. This needs credentials for the evaluator's provider.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Pre-Call Evaluation","lvl3":""}},
+{"objectID":"5db7566aedf90d539390892389df697a9ba96c4880f17fa840e40608ea86d19b","title":"Dynamic Guardrails","url":"/docs/features/guardrails#dynamic-guardrails","content":"A per-call middleware can carry a list you load at request time. It replaces the list from any default, because lists are replaced rather than combined.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Dynamic Guardrails","lvl3":""}},
+{"objectID":"1f76b8a28577849750e76dd464d6646aa9bdefe8523d12f311037f919fe24305","title":"API Reference","url":"/docs/features/guardrails#api-reference","content":"new NeuroLink({ middleware }): instance default, a MiddlewareFactoryOptions object\nconfigureGlobalMiddleware(options): process-wide default; replaces the previous one and throws a TypeError if options is not an object\nclearGlobalMiddleware(): remove the process-wide default\ngenerate({ middleware }), stream({ middleware }), generateText({ middleware }): per-call, highest precedence\npreset: \"security\": guardrails on, nothing configured\npreset: \"all\": analytics and guardrails\nmiddlewareConfig.guardrails: the configuration described above\n\nSee guardrails-ai-integration.md for pre-call evaluation in depth.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"API Reference","lvl3":""}},
+{"objectID":"85995a92035eb64e3e92878f16c818aa6b02432dfcdf8a2c1b7d0ec93efb0ceb","title":"Problem: Guardrails not filtering content","url":"/docs/features/guardrails#problem-guardrails-not-filtering-content","content":"Cause: One of the switches is missing, or only a preset is set. preset: \"security\" enables guardrails but gives them nothing to filter.\nSolution: Set all of these, on the call, the instance or the global default:\n\nIf a default is set but a call still passes through, check whether the call's own middleware switches guardrails off (disabledMiddleware or enabled: false) or replaces the list.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Problem: Guardrails not filtering content","lvl3":""}},
+{"objectID":"69697f2c3191cc2ac3f2d5a0a1e9295a7c6df79603d7a54068a65420afd69858","title":"Problem: Too many false positives (legitimate content filtered)","url":"/docs/features/guardrails#problem-too-many-false-positives-legitimate-content-filtered","content":"Cause: Terms match anywhere inside a word, so a short or common term catches unrelated words.\nSolution: Use specific terms, or a regular expression with word boundaries:","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Problem: Too many false positives (legitimate content filtered)","lvl3":""}},
 {"objectID":"eee2fe44ae3284cb63eec8fb3919e61415d44d01645db00a942193e14955f189","title":"Problem: A regex pattern is not filtering anything","url":"/docs/features/guardrails#problem-a-regex-pattern-is-not-filtering-anything","content":"Cause: The pattern was refused for its length or its shape, or it is not valid regex. Both are skipped, and the log says which\nSolution: Look for Refusing regex pattern (or Invalid regex pattern) in the error log. A refusal names the reason, for example a repeated group whose variable-length part can match the same text as what follows it. Separate the parts with a character they cannot match (([\\w-]+\\.)+ instead of ([\\w-]+)+), split a long pattern into several, or move plain terms to badWords.list","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Problem: A regex pattern is not filtering anything","lvl3":""}},
-{"objectID":"203e91cc171e445ffc3872091866a6a6da27809cb879259d7c595e54cf02b29f","title":"Problem: Model-based filter is slow","url":"/docs/features/guardrails#problem-model-based-filter-is-slow","content":"Cause: Using large/expensive model for filtering\nSolution:","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Problem: Model-based filter is slow","lvl3":""}},
-{"objectID":"3d6cce90a89ac6dc76fe42611ab3d172032a8879982db277948a0f6982c57923","title":"Problem: Guardrails not working in streaming mode","url":"/docs/features/guardrails#problem-guardrails-not-working-in-streaming-mode","content":"Cause: Streaming guardrails only support bad word filtering (not model-based)\nSolution:","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Problem: Guardrails not working in streaming mode","lvl3":""}},
-{"objectID":"95ef73f0ed68f3ce81ba9c25d863438aae71096384fc37cf1d5b3ee3e6219c69","title":"Content Filtering Strategy","url":"/docs/features/guardrails#content-filtering-strategy","content":"Start with presets - Use preset: \"security\" as baseline\nLayer protections - Combine bad words + model filtering\nUse lightweight filter models - gpt-4o-mini for speed\nTest thoroughly - Verify filtering doesn't break legitimate content\nMonitor and iterate - Track false positives/negatives","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Content Filtering Strategy","lvl3":""}},
-{"objectID":"a900aa88102d3e7f32205bee5665bb0732bca6cfe575bcdca0ca993a3590f28e","title":"Bad Word List Curation","url":"/docs/features/guardrails#bad-word-list-curation","content":"✅ Do:\nInclude specific harmful terms\nUse exact phrases, not single characters\nRegularly update based on user reports\nConsider context-specific terms for your domain\n\n❌ Don't:\nAdd common English words (high false positive rate)\nInclude single letters or short words\nRely solely on bad words (use model filter too)","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Bad Word List Curation","lvl3":""}},
-{"objectID":"8380babcbfc88033aac779943fa372013989ce1cd79ebbc9f1803af06cf2e27e","title":"Related Features","url":"/docs/features/guardrails#related-features","content":"HITL Workflows - User approval for risky actions\nMiddleware Architecture - Custom middleware development\nAnalytics Integration - Track filtered content metrics","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Related Features","lvl3":""}},
-{"objectID":"4dcd9d3f7dd025aac5e11ee4820832266eed4420348ba5f439bac3b9035737ae","title":"Migration Notes","url":"/docs/features/guardrails#migration-notes","content":"If upgrading from versions before v7.42.0:\nGuardrails are now enabled via middleware presets\nOld guardrailsConfig option deprecated - use middlewareConfig.guardrails\nNo breaking changes - existing configs still work\nRecommended: Switch to preset: \"security\" for simplified setup\n\nFor complete technical documentation and advanced integration patterns, see guardrails-ai-integration.md.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Migration Notes","lvl3":""}},
+{"objectID":"4678be93ebd07a49cf125f8b350254ac320ee6245be6e4d77746e5ebbebdee86","title":"Problem: Model-based filter is slow or does nothing","url":"/docs/features/guardrails#problem-model-based-filter-is-slow-or-does-nothing","content":"Cause: A large filter model adds a second round trip to every generate(). If the filter model cannot be created, for example a bare id that does not exist on the default provider, the failure is logged and the reply is returned unfiltered.\nSolution: Use a small model and name its provider:","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Problem: Model-based filter is slow or does nothing","lvl3":""}},
+{"objectID":"3d6cce90a89ac6dc76fe42611ab3d172032a8879982db277948a0f6982c57923","title":"Problem: Guardrails not working in streaming mode","url":"/docs/features/guardrails#problem-guardrails-not-working-in-streaming-mode","content":"Cause: stream() supports bad-word filtering and pre-call evaluation but not the model filter.\nSolution: Use badWords for streams, or use generate() when you need the model filter.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Problem: Guardrails not working in streaming mode","lvl3":""}},
+{"objectID":"54dd2e05fca81d0523db814f1329fd7f6c77c3bd34b65e43ef4b5afe5e0f7838","title":"Problem: NEUROLINK_MIDDLEWARE_PRESET has no effect","url":"/docs/features/guardrails#problem-neurolink_middleware_preset-has-no-effect","content":"Cause: The name is not default, all or security (it is then ignored, with a message on stderr), the variable was set after the NeuroLink instance was constructed, or a higher layer sets its own preset.\nSolution: Export it before the process starts, and check the layer order in Defaults and precedence.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Problem: NEUROLINK_MIDDLEWARE_PRESET has no effect","lvl3":""}},
+{"objectID":"a7e9e7181f3c4e419e083e152c274ddaf929b0240ee057ece97504c909ccc8f2","title":"Best Practices","url":"/docs/features/guardrails#best-practices","content":"Name what you filter. The presets enable guardrails; your word list or filter model does the filtering.\nSet a default for every call you cannot audit. An instance or global default is applied even where a call site forgot middleware.\nPrefer specific terms or word-boundary patterns to short common words.\nTest with a real reply. Run a call that should be filtered and one that should not, and check both.\nTreat the model filter as best effort. It fails open and sees only text parts.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Best Practices","lvl3":""}},
+{"objectID":"8380babcbfc88033aac779943fa372013989ce1cd79ebbc9f1803af06cf2e27e","title":"Related Features","url":"/docs/features/guardrails#related-features","content":"HITL Workflows - User approval for risky actions\nMiddleware Architecture - How the middleware chain is built\nBuilt-in Middleware - Analytics, guardrails and auto-evaluation\nGuardrails AI Integration - Pre-call evaluation and MiddlewareFactory usage","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Related Features","lvl3":""}},
+{"objectID":"4dcd9d3f7dd025aac5e11ee4820832266eed4420348ba5f439bac3b9035737ae","title":"Migration Notes","url":"/docs/features/guardrails#migration-notes","content":"Instance, global and environment defaults were added after guardrails were first released. They are optional: a call that passes its own middleware and sets no default behaves as it always did.","hierarchy":{"lvl0":"Features","lvl1":"Guardrails Middleware","lvl2":"Migration Notes","lvl3":""}},
 {"objectID":"4360e3d584ae161b59c6d6c3d6c7fe074f475883b020805602930215117fd0d7","title":"Human-in-the-Loop (HITL) Workflows","url":"/docs/features/hitl","content":"Human-in-the-Loop (HITL) Workflows\n\nSince: v7.39.0 | Status: Stable | Availability: SDK\n\nOverview\n\nWhat it does: HITL pauses AI tool execution to request explicit user approval before performing risky operations like deleting files, modifying databases, or making expensive API calls.\n\nWhy use it: Prevent costly mistakes and give users control over potentially dangerous AI actions. Think of it as an \"Are you sure?\" dialog for AI assistant operations.\n\nOnly use HITL for truly risky operations. Overusing confirmation prompts degrades user experience and can lead to \"confirmation fatigue\" where users approve actions without reading them.\n\nCommon use cases:\nFile deletion or modification operations\nDatabase write/delete operations\nExpensive third-party API calls\nIrreversible actions (sending emails, posting to social media)\nOperations accessing sensitive data\n\nQuick Start\n\nSDK Example\nTool identifier used by the AI to invoke this function\nDescribes tool purpose to the LLM for proper selection\nTriggers HITL checkpoint before execution\nActual implementation only runs after user approval\n\nHandling Confirmation in Your UI\n\nHITL uses an event-based workflow where the SDK emits confirmation requests and your app responds with user decisions.\nGet the event emitter - NeuroLink is not itself an emitter, so requests and responses both go through getEventEmitter()\nSubscribe to confirmation requests - NeuroLink emits one before running a tool that needs approval\nShow your confirmation UI with the tool name, arguments and remaining time\nReply on the same emitter, echoing the request's confirmationId so the response is matched to its request\nApproval decision determines if tool executes\nOptional: Handle cases where user doesn't respond in time\n\nConfiguration\n\n| Option                 | Type      | Default | Required | Description                          |\n| ---------------------- | --------- | ------- | -------- | ------------------------------------ |\n| requiresConfirmation | boolean | false | No       | Mark tool as requiring user approval |\n\nTool Registration\n\nHow It Works\n\nExecution Flow\nAI requests tool execution → Tool executor checks if tool requires confirmation\nConfirmation required? → Returns USER_CONFIRMATION_REQUIRED error to LLM\nLLM asks user → \"I need to [action]. Is that okay?\"\nUser responds:\nApprove → UI sets confirmation_received = true and retries tool execution\nDeny → UI sends \"User cancelled\" message back to LLM\nTool executes → Permission flag immediately resets to false\n\nSecurity Features\nOne-time permissions: Each approval works for exactly one action\nNo reuse: AI cannot reuse old permissions for new actions\nAutomatic reset: Permission flag clears immediately after use\nFail-safe: Defaults to requiring permission when in doubt\n\nAPI Reference\n\nEvent Types\n\nAll HITL events ride the SDK emitter returned by\nneurolink.getEventEmitter(); emitter below is that object.\n\nConfirmation Request Event (hitl:confirmation-request):\n\nConfirmation Response (emit from your app):\n\nTimeout Event (hitl:timeout):\n\nTroubleshooting\n\nProblem: Tool executes without asking for permission\n\nCause: Tool not marked with requiresConfirmation: true\nSolution:\nAdd this boolean flag to any tool that performs risky operations\n\nProblem: AI keeps asking for confirmation repeatedly\n\nCause: Confirmation responses not being sent or sent with wrong confirmationId\nSolution:\nExtract confirmation ID from the request event\nAlways respond to every confirmation request\nCritical: Use the same confirmationId from the request\n\nProblem: Confirmation dialog doesn't show\n\nCause: Not listening to hitl:confirmation-request event\nSolution:\nRegister the event handler early in your application startup\nAll subsequent tool executions will trigger confirmations when needed\n\nBest Practices\n\nStore user confirmation preferences to avoid repeated prompts for the same action type. For example, if a user approves \"delete temporary files\" once, cache that preference for similar low-risk deletions in the same session.\n\nFor Developers\nMark tools conservatively - If an operation could cause problems, require confirmation\nClear prompts - Ensure users understand exactly what will happen\nTest confirmation flow - Verify it works smoothly in your UI\nLog approvals - Keep audit trail of user decisions\nHandle denials gracefully - Allow users to try alternative approaches\n\nWhat to Mark as Requiring Confirmation\n\n✅ Do require confirmation:\nFile deletions\nDatabase writes/deletes\nSending emails or messages\nMaking purchases or payments\nModifying production systems\n\n❌ Don't require confirmation:\nRead-only operations\nAnswering questions\nGenerating content\nSearching/fetching data\n\nRelated Features\nGuardrails Middleware - Content filtering and safety checks\nCustom Tools - Building your own tools with HITL\nMiddleware Architecture - Advanced request interception\n\nMigration Notes\n\nIf upgrading from versions before v7.39.0:\nReview all existing tools for risk assessment\nAdd requiresConfirmation: tr","hierarchy":{"lvl0":"Features","lvl1":"Human-in-the-Loop (HITL) Workflows","lvl2":"","lvl3":""}},
 {"objectID":"d85931f065438b4475cf5fc7e7854d68dda75d45b96963229ea7accb7f456a86","title":"Human-in-the-Loop (HITL) Workflows","url":"/docs/features/hitl#human-in-the-loop-hitl-workflows","content":"Since: v7.39.0 | Status: Stable | Availability: SDK","hierarchy":{"lvl0":"Features","lvl1":"Human-in-the-Loop (HITL) Workflows","lvl2":"Human-in-the-Loop (HITL) Workflows","lvl3":""}},
 {"objectID":"f1d18bf2b10a871a2c1ba24625cfdba4ced0186612c32e21ff05afa4d8c614b5","title":"Overview","url":"/docs/features/hitl#overview","content":"What it does: HITL pauses AI tool execution to request explicit user approval before performing risky operations like deleting files, modifying databases, or making expensive API calls.\n\nWhy use it: Prevent costly mistakes and give users control over potentially dangerous AI actions. Think of it as an \"Are you sure?\" dialog for AI assistant operations.\n\nOnly use HITL for truly risky operations. Overusing confirmation prompts degrades user experience and can lead to \"confirmation fatigue\" where users approve actions without reading them.\n\nCommon use cases:\nFile deletion or modification operations\nDatabase write/delete operations\nExpensive third-party API calls\nIrreversible actions (sending emails, posting to social media)\nOperations accessing sensitive data","hierarchy":{"lvl0":"Features","lvl1":"Human-in-the-Loop (HITL) Workflows","lvl2":"Overview","lvl3":""}},
@@ -3584,9 +3588,9 @@
 {"objectID":"a6f80762fe08ca735f864be4541b6863065a0acfc5849344db2e40f38db28db3","title":"Verify file","url":"/docs/features/image-generation#verify-file","content":"file ./test-output/square.png","hierarchy":{"lvl0":"Features","lvl1":"Image Generation Streaming Guide","lvl2":"Verify file","lvl3":""}},
 {"objectID":"0de2f95f5125f67748b214f7ed674edcf44078cc415fa081f6edf46428d7c65f","title":"Output: ./test-output/square.png: PNG image data, 1024 x 1024, 8-bit/color RGB","url":"/docs/features/image-generation#output-test-outputsquarepng-png-image-data-1024-x-1024-8-bitcolor-rgb","content":"`","hierarchy":{"lvl0":"Features","lvl1":"Image Generation Streaming Guide","lvl2":"Output: ./test-output/square.png: PNG image data, 1024 x 1024, 8-bit/color RGB","lvl3":""}},
 {"objectID":"aa931e13d4db435aa9de30f72f599057c2977804e4316c3cb61f84d8e43fd97e","title":"Conclusion","url":"/docs/features/image-generation#conclusion","content":"NeuroLink's image generation streaming provides a unified interface for both text and image generation. The fake streaming approach ensures consistency while maintaining the benefits of streaming APIs. By following the patterns and examples in this guide, you can effectively integrate image generation into your applications.\n\nFor more information:\nAPI Reference\nProvider Comparison\nProvider Status Monitoring","hierarchy":{"lvl0":"Features","lvl1":"Image Generation Streaming Guide","lvl2":"Conclusion","lvl3":""}},
-{"objectID":"e3366024d1a12d0c73fbcb1592927d2dcc0d92c12eaea661cf5164bb7217c32f","title":"Feature Guides","url":"/docs/features","content":"Feature Guides\n\nComprehensive guides for all NeuroLink features organized by category. Each guide includes setup, usage patterns, configuration, and troubleshooting.\n\nLatest Features (Q1 2026)\n\n| Feature                                                                                  | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |\n| ---------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |\n| decide Inference Type                      | A third inference type alongside generate/stream: typed, calibrated boolean/choice/score judgments in one parallel pass — no text. ~400ms, ~$0.00002 per decision on TypeSafe; Perplexity adds about 65 ms for each further question, up to 128 (see its guide); Cloudflare Clef takes up to 64 questions and about 2,048 tokens of state (see its guide). Providers: TypeSafe Jev; Laya, open-weights and self-hosted; XOR, Juspay's open-weights model that also reads images and video; SubQ, a hosted API; Perplexity, a hosted API that also reads images; Cloudflare Clef, hosted on Workers AI, which also reads images; and OpenAI Decisions, OpenAI's Decisions API (public beta), which also reads images. Fail-open: a no-op when no decision provider is configured (the credentials of Perplexity, Cloudflare Clef and OpenAI Decisions are shared with their text providers). |\n| Model Routing with a Decision Model | One round trip answers difficulty, capabilities, risk, context scope and the model pick, with asymmetric confidence bars (0.3 up / 0.6 down).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            ","hierarchy":{"lvl0":"Features","lvl1":"Feature Guides","lvl2":"","lvl3":""}},
+{"objectID":"e3366024d1a12d0c73fbcb1592927d2dcc0d92c12eaea661cf5164bb7217c32f","title":"Feature Guides","url":"/docs/features","content":"Feature Guides\n\nComprehensive guides for all NeuroLink features organized by category. Each guide includes setup, usage patterns, configuration, and troubleshooting.\n\nLatest Features (Q1 2026)\n\n| Feature                                                                                  | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |\n| ---------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| decide Inference Type                      | A third inference type alongside generate/stream: typed, calibrated boolean/choice/score judgments in one parallel pass — no text. ~400ms, ~$0.00002 per decision on TypeSafe; Perplexity adds about 65 ms for each further question, up to 128 (see its guide); Cloudflare Clef takes up to 64 questions and about 2,048 tokens of state (see its guide). Providers: TypeSafe Jev; Laya, open-weights and self-hosted; XOR, Juspay's open-weights model that also reads images and video; SubQ, a hosted API; Perplexity, a hosted API that also reads images; Cloudflare Clef, hosted on Workers AI, which also reads images; OpenAI Decisions, OpenAI's Decisions API (public beta), which also reads images; and ScaleDown, a hosted API that reads text only. Fail-open: a no-op when no decision provider is configured (the credentials of Perplexity, Cloudflare Clef and OpenAI Decisions are shared with their text providers; ScaleDown's is its own). |\n| Model Routing with a Decision Model | One round trip answers difficulty, capabilities, risk, context scope and the model pick, with asymmetric confidence bars (0.3 up / 0.6 down).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                ","hierarchy":{"lvl0":"Features","lvl1":"Feature Guides","lvl2":"","lvl3":""}},
 {"objectID":"a477c76a577a17e7aeaf89736289a0a41734b4bcfe479a523cdd9e2051bd0ba7","title":"Feature Guides","url":"/docs/features#feature-guides","content":"Comprehensive guides for all NeuroLink features organized by category. Each guide includes setup, usage patterns, configuration, and troubleshooting.","hierarchy":{"lvl0":"Features","lvl1":"Feature Guides","lvl2":"Feature Guides","lvl3":""}},
-{"objectID":"a43e81fa4a6c594f8b6b4cfc216db1356d75b3e4ebb54279dd2da7f69001d28f","title":"Latest Features (Q1 2026)","url":"/docs/features#latest-features-q1-2026","content":"| Feature                                                                                  | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |\n| ---------------------------------------------------------------------------------------- |","hierarchy":{"lvl0":"Features","lvl1":"Feature Guides","lvl2":"Latest Features (Q1 2026)","lvl3":""}},
+{"objectID":"a43e81fa4a6c594f8b6b4cfc216db1356d75b3e4ebb54279dd2da7f69001d28f","title":"Latest Features (Q1 2026)","url":"/docs/features#latest-features-q1-2026","content":"| Feature                                                                                  | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |\n| ---------------------------------------------------------------------------------------- |","hierarchy":{"lvl0":"Features","lvl1":"Feature Guides","lvl2":"Latest Features (Q1 2026)","lvl3":""}},
 {"objectID":"84c604ce940f85fbe8aa8d94675e4c1dae80b22ede558d58102426598f6a7d06","title":"Core Features (shipped 2025)","url":"/docs/features#core-features-shipped-2025","content":"| Feature                                                  | Description                                                                                           |\n| -------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- |\n| Image Generation | Generate images from text prompts using Gemini models via Vertex AI or Google AI Studio.              |\n| Enterprise HITL                | Production-ready HITL with approval workflows, confidence thresholds, and enterprise patterns.        |\n| Interactive CLI                | AI development environment with loop mode, session variables, and conversation memory.                |\n| MCP Tools Showcase          | Complete guide to 6 built-in tools and connecting external MCP servers across 6 categories.           |\n| Human-in-the-Loop (HITL)                  | Pause AI tool execution for user approval before risky operations like file deletion or API calls.    |\n| Guardrails Middleware               | Content filtering, PII detection, and safety checks for AI outputs with zero configuration.           |\n| Redis Conversation Export | Export complete session history as JSON for analytics, debugging, and compliance auditing.            |\n| Context Compaction          | Automatic conversation compression for long-running sessions to stay within token limits.             |\n| LiteLLM Integration     | Access 100+ AI models across a broad range of AI providers through unified LiteLLM routing interface. |\n| SageMaker Integration | Deploy and use custom-trained models on AWS SageMaker infrastructure with full control.               |","hierarchy":{"lvl0":"Features","lvl1":"Feature Guides","lvl2":"Core Features (shipped 2025)","lvl3":""}},
 {"objectID":"f64aba3e564a8f56064839c33b533971468da990a9cc95d86ceb1e9d13223478","title":"Earlier Core Features (shipped Q3 2025)","url":"/docs/features#earlier-core-features-shipped-q3-2025","content":"| Feature                                                       | Description                                                                                             |\n| ------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------- |\n| Multimodal Chat Experiences         | Stream text and images together with automatic provider fallbacks and format conversion.                |\n| CSV File Support                        | Process CSV files for data analysis with automatic format conversion. Works across supported providers. |\n| PDF File Support                        | Process PDF documents for visual analysis and content extraction. Native provider support.              |\n| Office Documents                   | Process DOCX, PPTX, XLSX files for document analysis. Native Bedrock, Vertex, Anthropic support.        |\n| Auto Evaluation Engine              | Automated quality scoring and metrics export for AI response validation using LLM-as-judge.             |\n| CLI Loop Sessions                 | Persistent interactive mode with conversation memory and session state for prompt engineering.          |\n| Regional Streaming Controls      | Region-specific model deployment and routing for compliance and latency optimization.                   |\n| Provider Orchestration Brain | Adaptive provider and model selection with intelligent fallbacks based on task classification.          |","hierarchy":{"lvl0":"Features","lvl1":"Feature Guides","lvl2":"Earlier Core Features (shipped Q3 2025)","lvl3":""}},
 {"objectID":"fbd682fc148477639d38ab175c38363d003a1464834e84b2a48d88087d001b91","title":"Platform Capabilities at a Glance","url":"/docs/features#platform-capabilities-at-a-glance","content":"| Category                 | Features                                                                                                           | Documentation                                                                                                                            |\n| ------------------------ | ------------------------------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------- |\n| Provider unification | Provider integrations with automatic failover, cost-aware routing, providerFallback policy, modelChain config  | Provider Setup                                                                                   |\n| Multimodal pipeline  | Stream images + CSV data + PDF documents + Office files across providers with auto-detection for mixed file types. | Multimodal Guide, CSV Support, PDF Support, Office Docs |\n| Voice pipeline       | TTS (6 providers) + STT (4 providers) + realtime APIs (OpenAI Realtime, Gemini Live)                               | TTS Guide, STT Guide, Realtime Services                                 |\n| Quality & governance | Auto-evaluation engine (14 scorers), guardrails middleware, HITL workflows, audit logging                          | Auto Evaluation, Guardrails, HITL                                                      |\n| Memory & context     | Per-user condensed memory (S3/Redis/SQLite), Redis session export, 5-stage context compaction                      | Conversation Memory, Memory, Redis Export                           |\n| CLI tooling          | Loop sessions, setup wizard, config validation, Redis auto-detect, JSON output, TTS/STT flags                      | CLI Loop, CLI Commands                                                                     |\n| Enterprise ops       | Claude proxy, OTLP observability, OpenObserve dashboard, regional routing,","hierarchy":{"lvl0":"Features","lvl1":"Feature Guides","lvl2":"Platform Capabilities at a Glance","lvl3":""}},
@@ -4518,8 +4522,8 @@
 {"objectID":"5594cc075cbacf6979c75d68f20b45ccff0513a1e4c7087d756491900cbb1c4d","title":"Model Detection Utilities","url":"/docs/features/thinking-configuration#model-detection-utilities","content":"NeuroLink provides utilities to check thinking support:","hierarchy":{"lvl0":"Features","lvl1":"Extended Thinking Configuration","lvl2":"Model Detection Utilities","lvl3":""}},
 {"objectID":"559dd274be4e9bdfeabf680692f3830f92a5c13caefc42acfb819d9635f2dad7","title":"Important Notes","url":"/docs/features/thinking-configuration#important-notes","content":"Provider compatibility: Thinking configuration is provider-specific. Gemini uses thinkingLevel, Claude uses budgetTokens\nToken consumption: Extended thinking uses additional tokens beyond the response\nLatency impact: Higher thinking levels increase response time\nNot all models support thinking: Check supportsThinkingConfig() before enabling\nStreaming support: Thinking configuration works with both generate() and stream()","hierarchy":{"lvl0":"Features","lvl1":"Extended Thinking Configuration","lvl2":"Important Notes","lvl3":""}},
 {"objectID":"4bd3d62cf46ca16e0b9a52768c51ca4bced1f07f6c32f9a141c0cbfc599bcefb","title":"See Also","url":"/docs/features/thinking-configuration#see-also","content":"API Reference\nProvider Configuration\nStreaming","hierarchy":{"lvl0":"Features","lvl1":"Extended Thinking Configuration","lvl2":"See Also","lvl3":""}},
-{"objectID":"ffa6f024a99609da8db45d9a02bc76178d7fac894f00af93b830f9b356e86f3c","title":"Tool / MCP routing by decision model","url":"/docs/features/tool-routing-decision-model","content":"Tool / MCP routing by decision model\n\nThe shipped tool router asks a generative model for {servers: string[]} on\na 15-second budget. That shape cannot express uncertainty — a server is in\nthe list or it is not, and the only recourse for a model that is unsure is to\ninclude it. A decision model instead\nasks one calibrated yes/no question per server, in a single round trip of\nabout 400ms and $0.00002 on TypeSafe (on Perplexity each further question adds\nabout 65 ms, up to 128 in a request; see\nits guide; on\nCloudflare Clef a request takes up to 64 questions, and tryDecide() splits a\nlarger set into batches of 64; see\nits guide; OpenAI Decisions\ndeclares no question cap, so a set goes whole and what the endpoint accepts is\nnot established, see\nits guide), and each\nanswer comes back with a real probability rather than a name that either did or\ndidn't make a list.\n\nThe degradation contract. selectServersByDecision() is used only when a\ndecision provider is configured; hosts never wire decideFn by hand — it is\nbound automatically wherever tool routing resolves, using the same\ntryDecide() that returns null on any failure or absent configuration. No\ndecision provider, a failed call, fewer than two candidate servers, or an\nanswer set that would drop nothing — any of these fall straight through to\nthe existing generative router, unchanged.\n\nOne question per server\n\nEach routable server gets its own boolean question, built from its\ndescription (or, absent one, its tool names):\n\nA server is excluded only on a confident false — minDropConfidence\ndefaults to 0.6. undefined (unanswered, malformed, or too close to a\ncoin flip) and a confident true both keep the server. This is the same\nasymmetry as the classifier's upgrade/downgrade bars: keeping an unneeded\nserver costs a few hundred tokens of tool definitions; dropping a needed one\nbreaks the turn outright, because the model can never call a tool it was\nnever shown and has no way to ask for it back.\n\nTwo size guards bound the request: at most 200 servers are asked about\nin one batch (MAX_SERVERS), and the query text sent as state is capped at\n10,000 characters. Servers past the cap are never asked about and are\ntherefore always kept — a server that was not offered to the model must\nnever be silently dropped by its own absence from the question set.\n\nNeuroLink refuses a state it estimates at more than 16,000 tokens with\nmax_tokens_exceeded. A query long enough to be estimated above that is refused,\nand routing falls through to the generative router, as it does on any failure.\n\nThe wording that made this work: a measured A/B result\n\nThe first phrasing tried was the obvious one: \"answering this request will\nrequire calling at least one tool from this server,\" with false meaning\n\"this server is unrelated, OR the request needs no tool at all.\" Measured\nagainst a 10-request × 5-server labelled set, it separated correctly but\nweakly — unrelated servers averaged p = 0.31 and reached as high as\n0.80, so at the 0.6 drop bar only 12 of 39 unneeded servers were\nactually dropped.\n\nThree changes fixed it: naming the server explicitly, asking in the present\ntense about what carrying out the request involves rather than what\n\"will require,\" and splitting the bundled false criterion (which was\nreally two separate claims joined by \"or\") into one single claim. That\nmoved unrelated servers to a mean of p = 0.03 with a maximum of 0.35\n— 37 of 39 dropped at the same 0.6 bar, still with zero wrong drops.\n\nThe lesson generalises past this one question: a decision model reads\nliterally, and an \"or\" in a criterion is two questions wearing one coat. Each\nhalf of a compound criterion pulls the answer toward the middle whenever\nonly one half is true, which is exactly the muddy, hard-to-gate signal the\nfirst version produced.\n\nWhat this is bad at\nIt reasons about servers, not individual tools. The unit of decision is\n  a whole MCP server; a server with twenty tools where the request needs one\n  is kept or dropped as a unit, not tool-by-tool.\nThe description quality bounds the question quality. A server with no\n  declared description falls back to a comma-joined list of its own tool\n  names, which carries much less signal than a well-written one-line\n  description — the wording fix above only helps once the server's own text\n  is legible to a literal reader.\nA close call still resolves to \"keep.\" There is no partial exclusion;\n  anything from a coin flip up to just under 0.6 confidence is treated\n  identically to a confident true.\nIt shares the base model's general limits — literal reading, no\n  arithmetic, accuracy sensitive to a noisy or oversized state — all\n  described in\n  what decide is bad at.\nThe query is untrusted input sent as state, and this module does not\n  sanitize it. The reach is deliberately bounded, not zero. Server ids are\n  never read back off the wire (answers are matched by position, not by name),\n  so a crafted query cannot name a tool that was not already registered, and\n  the ","hierarchy":{"lvl0":"Features","lvl1":"Tool / MCP routing by decision model","lvl2":"","lvl3":""}},
-{"objectID":"4b84d2a2ea7a8b34624599647a02792ace2fc30650e5bb5d3da1ea68e46fdad8","title":"Tool / MCP routing by decision model","url":"/docs/features/tool-routing-decision-model#tool-mcp-routing-by-decision-model","content":"The shipped tool router asks a generative model for {servers: string[]} on\na 15-second budget. That shape cannot express uncertainty — a server is in\nthe list or it is not, and the only recourse for a model that is unsure is to\ninclude it. A decision model instead\nasks one calibrated yes/no question per server, in a single round trip of\nabout 400ms and $0.00002 on TypeSafe (on Perplexity each further question adds\nabout 65 ms, up to 128 in a request; see\nits guide; on\nCloudflare Clef a request takes up to 64 questions, and tryDecide() splits a\nlarger set into batches of 64; see\nits guide; OpenAI Decisions\ndeclares no question cap, so a set goes whole and what the endpoint accepts is\nnot established, see\nits guide), and each\nanswer comes back with a real probability rather than a name that either did or\ndidn't make a list.\n\nThe degradation contract. selectServersByDecision() is used only when a\ndecision provider is configured; hosts never wire decideFn by hand — it is\nbound automatically wherever tool routing resolves, using the same\ntryDecide() that returns null on any failure or absent configuration. No\ndecision provider, a failed call, fewer than two candidate servers, or an\nanswer set that would drop nothing — any of these fall straight through to\nthe existing generative router, unchanged.","hierarchy":{"lvl0":"Features","lvl1":"Tool / MCP routing by decision model","lvl2":"Tool / MCP routing by decision model","lvl3":""}},
+{"objectID":"ffa6f024a99609da8db45d9a02bc76178d7fac894f00af93b830f9b356e86f3c","title":"Tool / MCP routing by decision model","url":"/docs/features/tool-routing-decision-model","content":"Tool / MCP routing by decision model\n\nThe shipped tool router asks a generative model for {servers: string[]} on\na 15-second budget. That shape cannot express uncertainty — a server is in\nthe list or it is not, and the only recourse for a model that is unsure is to\ninclude it. A decision model instead\nasks one calibrated yes/no question per server, in a single round trip of\nabout 400ms and $0.00002 on TypeSafe (on Perplexity each further question adds\nabout 65 ms, up to 128 in a request; see\nits guide; on\nCloudflare Clef a request takes up to 64 questions, and tryDecide() splits a\nlarger set into batches of 64; see\nits guide; OpenAI Decisions\ndeclares no question cap, so a set goes whole and what the endpoint accepts is\nnot established, see\nits guide;\nScaleDown declares none either, and its maximum is not measured, see\nits guide), and each\nanswer comes back with a real probability rather than a name that either did or\ndidn't make a list.\n\nThe degradation contract. selectServersByDecision() is used only when a\ndecision provider is configured; hosts never wire decideFn by hand — it is\nbound automatically wherever tool routing resolves, using the same\ntryDecide() that returns null on any failure or absent configuration. No\ndecision provider, a failed call, fewer than two candidate servers, or an\nanswer set that would drop nothing — any of these fall straight through to\nthe existing generative router, unchanged.\n\nOne question per server\n\nEach routable server gets its own boolean question, built from its\ndescription (or, absent one, its tool names):\n\nA server is excluded only on a confident false — minDropConfidence\ndefaults to 0.6. undefined (unanswered, malformed, or too close to a\ncoin flip) and a confident true both keep the server. This is the same\nasymmetry as the classifier's upgrade/downgrade bars: keeping an unneeded\nserver costs a few hundred tokens of tool definitions; dropping a needed one\nbreaks the turn outright, because the model can never call a tool it was\nnever shown and has no way to ask for it back.\n\nTwo size guards bound the request: at most 200 servers are asked about\nin one batch (MAX_SERVERS), and the query text sent as state is capped at\n10,000 characters. Servers past the cap are never asked about and are\ntherefore always kept — a server that was not offered to the model must\nnever be silently dropped by its own absence from the question set.\n\nNeuroLink refuses a state it estimates at more than 16,000 tokens with\nmax_tokens_exceeded. A query long enough to be estimated above that is refused,\nand routing falls through to the generative router, as it does on any failure.\n\nThe wording that made this work: a measured A/B result\n\nThe first phrasing tried was the obvious one: \"answering this request will\nrequire calling at least one tool from this server,\" with false meaning\n\"this server is unrelated, OR the request needs no tool at all.\" Measured\nagainst a 10-request × 5-server labelled set, it separated correctly but\nweakly — unrelated servers averaged p = 0.31 and reached as high as\n0.80, so at the 0.6 drop bar only 12 of 39 unneeded servers were\nactually dropped.\n\nThree changes fixed it: naming the server explicitly, asking in the present\ntense about what carrying out the request involves rather than what\n\"will require,\" and splitting the bundled false criterion (which was\nreally two separate claims joined by \"or\") into one single claim. That\nmoved unrelated servers to a mean of p = 0.03 with a maximum of 0.35\n— 37 of 39 dropped at the same 0.6 bar, still with zero wrong drops.\n\nThe lesson generalises past this one question: a decision model reads\nliterally, and an \"or\" in a criterion is two questions wearing one coat. Each\nhalf of a compound criterion pulls the answer toward the middle whenever\nonly one half is true, which is exactly the muddy, hard-to-gate signal the\nfirst version produced.\n\nWhat this is bad at\nIt reasons about servers, not individual tools. The unit of decision is\n  a whole MCP server; a server with twenty tools where the request needs one\n  is kept or dropped as a unit, not tool-by-tool.\nThe description quality bounds the question quality. A server with no\n  declared description falls back to a comma-joined list of its own tool\n  names, which carries much less signal than a well-written one-line\n  description — the wording fix above only helps once the server's own text\n  is legible to a literal reader.\nA close call still resolves to \"keep.\" There is no partial exclusion;\n  anything from a coin flip up to just under 0.6 confidence is treated\n  identically to a confident true.\nIt shares the base model's general limits — literal reading, no\n  arithmetic, accuracy sensitive to a noisy or oversized state — all\n  described in\n  what decide is bad at.\nThe query is untrusted input sent as state, and this module does not\n  sanitize it. The reach is deliberately bounded, not zero. Server ids are\n  never read back off the wire (answers are matched by position, not by name),\n  s","hierarchy":{"lvl0":"Features","lvl1":"Tool / MCP routing by decision model","lvl2":"","lvl3":""}},
+{"objectID":"4b84d2a2ea7a8b34624599647a02792ace2fc30650e5bb5d3da1ea68e46fdad8","title":"Tool / MCP routing by decision model","url":"/docs/features/tool-routing-decision-model#tool-mcp-routing-by-decision-model","content":"The shipped tool router asks a generative model for {servers: string[]} on\na 15-second budget. That shape cannot express uncertainty — a server is in\nthe list or it is not, and the only recourse for a model that is unsure is to\ninclude it. A decision model instead\nasks one calibrated yes/no question per server, in a single round trip of\nabout 400ms and $0.00002 on TypeSafe (on Perplexity each further question adds\nabout 65 ms, up to 128 in a request; see\nits guide; on\nCloudflare Clef a request takes up to 64 questions, and tryDecide() splits a\nlarger set into batches of 64; see\nits guide; OpenAI Decisions\ndeclares no question cap, so a set goes whole and what the endpoint accepts is\nnot established, see\nits guide;\nScaleDown declares none either, and its maximum is not measured, see\nits guide), and each\nanswer comes back with a real probability rather than a name that either did or\ndidn't make a list.\n\nThe degradation contract. selectServersByDecision() is used only when a\ndecision provider is configured; hosts never wire decideFn by hand — it is\nbound automatically wherever tool routing resolves, using the same\ntryDecide() that returns null on any failure or absent configuration. No\ndecision provider, a failed call, fewer than two candidate servers, or an\nanswer set that would drop nothing — any of these fall straight through to\nthe existing generative router, unchanged.","hierarchy":{"lvl0":"Features","lvl1":"Tool / MCP routing by decision model","lvl2":"Tool / MCP routing by decision model","lvl3":""}},
 {"objectID":"0b9203bc3ddca1d61b578f4bea25de508f6a241da96f270ca4742163bd99884a","title":"One question per server","url":"/docs/features/tool-routing-decision-model#one-question-per-server","content":"Each routable server gets its own boolean question, built from its\ndescription (or, absent one, its tool names):\n\nA server is excluded only on a confident false — minDropConfidence\ndefaults to 0.6. undefined (unanswered, malformed, or too close to a\ncoin flip) and a confident true both keep the server. This is the same\nasymmetry as the classifier's upgrade/downgrade bars: keeping an unneeded\nserver costs a few hundred tokens of tool definitions; dropping a needed one\nbreaks the turn outright, because the model can never call a tool it was\nnever shown and has no way to ask for it back.\n\nTwo size guards bound the request: at most 200 servers are asked about\nin one batch (MAX_SERVERS), and the query text sent as state is capped at\n10,000 characters. Servers past the cap are never asked about and are\ntherefore always kept — a server that was not offered to the model must\nnever be silently dropped by its own absence from the question set.\n\nNeuroLink refuses a state it estimates at more than 16,000 tokens with\nmax_tokens_exceeded. A query long enough to be estimated above that is refused,\nand routing falls through to the generative router, as it does on any failure.","hierarchy":{"lvl0":"Features","lvl1":"Tool / MCP routing by decision model","lvl2":"One question per server","lvl3":""}},
 {"objectID":"19352c9bf82d208bbd76e4303e3e35ac7d4f7497562a4f1ee006d8cf1fae5f6b","title":"The wording that made this work: a measured A/B result","url":"/docs/features/tool-routing-decision-model#the-wording-that-made-this-work-a-measured-ab-result","content":"The first phrasing tried was the obvious one: \"answering this request will\nrequire calling at least one tool from this server,\" with false meaning\n\"this server is unrelated, OR the request needs no tool at all.\" Measured\nagainst a 10-request × 5-server labelled set, it separated correctly but\nweakly — unrelated servers averaged p = 0.31 and reached as high as\n0.80, so at the 0.6 drop bar only 12 of 39 unneeded servers were\nactually dropped.\n\nThree changes fixed it: naming the server explicitly, asking in the present\ntense about what carrying out the request involves rather than what\n\"will require,\" and splitting the bundled false criterion (which was\nreally two separate claims joined by \"or\") into one single claim. That\nmoved unrelated servers to a mean of p = 0.03 with a maximum of 0.35\n— 37 of 39 dropped at the same 0.6 bar, still with zero wrong drops.\n\nThe lesson generalises past this one question: a decision model reads\nliterally, and an \"or\" in a criterion is two questions wearing one coat. Each\nhalf of a compound criterion pulls the answer toward the middle whenever\nonly one half is true, which is exactly the muddy, hard-to-gate signal the\nfirst version produced.","hierarchy":{"lvl0":"Features","lvl1":"Tool / MCP routing by decision model","lvl2":"The wording that made this work: a measured A/B result","lvl3":""}},
 {"objectID":"9c9af1c1ff58051674501cb10c371fdcf50758c935a2dfa168df9719969bb7af","title":"What this is bad at","url":"/docs/features/tool-routing-decision-model#what-this-is-bad-at","content":"It reasons about servers, not individual tools. The unit of decision is\n  a whole MCP server; a server with twenty tools where the request needs one\n  is kept or dropped as a unit, not tool-by-tool.\nThe description quality bounds the question quality. A server with no\n  declared description falls back to a comma-joined list of its own tool\n  names, which carries much less signal than a well-written one-line\n  description — the wording fix above only helps once the server's own text\n  is legible to a literal reader.\nA close call still resolves to \"keep.\" There is no partial exclusion;\n  anything from a coin flip up to just under 0.6 confidence is treated\n  identically to a confident true.\nIt shares the base model's general limits — literal reading, no\n  arithmetic, accuracy sensitive to a noisy or oversized state — all\n  described in\n  what decide is bad at.\nThe query is untrusted input sent as state, and this module does not\n  sanitize it. The reach is deliberately bounded, not zero. Server ids are\n  never read back off the wire (answers are matched by position, not by name),\n  so a crafted query cannot name a tool that was not already registered, and\n  the result is only a denylist appended to excludeTools. It can still steer\n  a verdict: a confident false (0.6 by default) withholds that server's\n  tools for the turn. Do not rely on this router as a security control; see\n  the Untrusted input bullet under\n  what decide is bad at.","hierarchy":{"lvl0":"Features","lvl1":"Tool / MCP routing by decision model","lvl2":"What this is bad at","lvl3":""}},
@@ -4531,7 +4535,7 @@
 {"objectID":"25734eb9746d69fee91c94beaab37ba0433ce9d1bc944c78ac54b14ddaf10f83","title":"Result","url":"/docs/features/transcribe#result","content":"TranscribeResult is the engine's STTResult (text, confidence,\nlanguage, duration, words, segments, speakers, metadata) plus what\nthe layers did:\n\n| Field              | Meaning                                                                                                  |\n| ------------------ | -------------------------------------------------------------------------------------------------------- |\n| raw              | The engine's text before any correction.                                                                 |\n| text             | The final text — the corrected text when correction changed it, else raw.                              |\n| corrected        | Present only when correction changed the text.                                                           |\n| languageDetected | false when the engine reported a low language-detection score.                                         |\n| languageScores   | Candidate languages with the engine's own scores, where it reports them.                                 |\n| decisions        | One record per dictionary candidate: what was heard, the term, term or literal, and the probability. |\n| engine           | { provider, model, fallbackUsed?, secondOpinion? }.                                                    |\n| timings          | transcribeMs, secondOpinionMs, decideMs, rewriteMs, correctionMs, totalMs.                   |\n| steps            | A readable trail: \"dictionary 2/2\", \"rewrite timed out · kept text\" …                                |","hierarchy":{"lvl0":"Features","lvl1":"The transcribe inference type","lvl2":"Result","lvl3":""}},
 {"objectID":"cfe99c9441a998e039ecbe92d247fb981326eb5e5a56e51aa7abb49fd4662663","title":"Providers and environment","url":"/docs/features/transcribe#providers-and-environment","content":"| Provider (aliases)                            | Configured by                                                                               | Notes                                                   |\n| --------------------------------------------- | ------------------------------------------------------------------------------------------- | ------------------------------------------------------- |\n| whisper (openai-stt, openai-compatible) | OPENAI_STT_API_KEY or OPENAI_API_KEY; OPENAI_STT_BASE_URL for another server          | Any OpenAI-compatible /audio/transcriptions endpoint. |\n| deepgram                                    | DEEPGRAM_API_KEY                                                                          | Native streaming.                                       |\n| elevenlabs-stt (scribe, elevenlabs)     | ELEVENLABS_API_KEY                                                                        |                                                         |\n| google-stt                                  | GOOGLE_API_KEY, GOOGLE_AI_API_KEY, GEMINI_API_KEY or GOOGLE_APPLICATION_CREDENTIALS |                                                         |\n| azure-stt                                   | AZURE_SPEECH_KEY                                                                          |                                                         |\n| whistle (local)                           | nothing                                                                                     | Built-in, local, always available.                      |\n\nThe order of that table is precedence: without provider or\nNEUROLINK_STT_PROVIDER, the first configured provider is used, and Whistle\nonly when none is.","hierarchy":{"lvl0":"Features","lvl1":"The transcribe inference type","lvl2":"Providers and environment","lvl3":""}},
 {"objectID":"1c83e72d0998d73b45f4a810514691aa3fb8979166eea4883ffe4bec728140d7","title":"Dictionary","url":"/docs/features/transcribe#dictionary","content":"A dictionary teaches the pipeline the words an engine cannot know — product\nnames, people's names, internal jargon — and how engines tend to mis-hear them:\n\n| Field     | Required | Use                                                                                               |\n| --------- | -------- | ------------------------------------------------------------------------------------------------- |\n| term    | yes      | The correct spelling, written exactly like this in the output.                                    |\n| heardAs | no       | Mis-hearings, in any script. Each one found in a transcript becomes a candidate for replacement.  |\n| meaning | no       | What the term means, given to the decision model and the rewrite so they can tell term from word. |\n\nThe dictionary is used three ways: its terms are sent to engines that take a\nprompt or vocabulary, each heardAs found in the transcript becomes a\ncandidate repair, and the whole list is given to the rewrite model.","hierarchy":{"lvl0":"Features","lvl1":"The transcribe inference type","lvl2":"Dictionary","lvl3":""}},
-{"objectID":"bb0871357646359062274d9ec5e47b41771d77b2c86ccd2d29cf96f43a538b64","title":"Correction","url":"/docs/features/transcribe#correction","content":"Correction runs on a finished transcript (or on each final utterance when\nstreaming), in two stages:\nDictionary, guarded. Each candidate is put to the decision model as one\n   question — did the speaker mean the term, or the ordinary word? — so a\n   product name never rewrites the everyday word it sounds like. All\n   candidates go in one tryDecide() call. With no decision provider, on a\n   timeout (decideTimeoutMs, default 4000) or on any error, the dictionary is\n   applied as is. guard: \"none\" always applies it. The guard asks for the\n   default decision provider, and it sends the first 4,000 characters of the\n   transcript to it. A lone OPENAI_API_KEY counts as one (it is also Whisper's\n   key), so a host that set it only for Whisper and passes a dictionary sends the\n   transcript text to OpenAI's Decisions endpoint, unless OPENAI_BASE_URL,\n   OPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com (a gateway's\n   key is not used for the default); see\n   One key, two providers.\nRewrite. A text model fixes spelling, punctuation and capitalisation,\n   writes dictionary terms correctly, reconciles a second opinion and writes\n   Indic-script words in Latin letters the way people type them (Hinglish,\n   Tanglish …) — never translating them. transliterate: \"native\" keeps the\n   original script. rewrite: false stops after the dictionary.\n\n| Option            | Default                                                                             | Meaning                                                                    |\n| ----------------- | ----------------------------------------------------------------------------------- | -------------------------------------------------------------------------- |\n| enabled         | on with a dictionary or any correction object, unless it is only rewrite: false | Run the layer at all.","hierarchy":{"lvl0":"Features","lvl1":"The transcribe inference type","lvl2":"Correction","lvl3":""}},
+{"objectID":"bb0871357646359062274d9ec5e47b41771d77b2c86ccd2d29cf96f43a538b64","title":"Correction","url":"/docs/features/transcribe#correction","content":"Correction runs on a finished transcript (or on each final utterance when\nstreaming), in two stages:\nDictionary, guarded. Each candidate is put to the decision model as one\n   question — did the speaker mean the term, or the ordinary word? — so a\n   product name never rewrites the everyday word it sounds like. All\n   candidates go in one tryDecide() call. With no decision provider, on a\n   timeout (decideTimeoutMs, default 4000) or on any error, the dictionary is\n   applied as is. guard: \"none\" always applies it. The guard asks for the\n   default decision provider, and it sends the first 4,000 characters of the\n   transcript to it. A lone OPENAI_API_KEY counts as one (it is also Whisper's\n   key), so a host that set it only for Whisper and passes a dictionary sends the\n   transcript text to OpenAI's Decisions endpoint, unless OPENAI_BASE_URL,\n   OPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com (a gateway's\n   key is not used for the default); see\n   One key, two providers.\n   SCALEDOWN_API_KEY, which no other provider reads, counts too: with it as the\n   only decision key, the transcript text goes to ScaleDown, the last provider in\n   the order.\nRewrite. A text model fixes spelling, punctuation and capitalisation,\n   writes dictionary terms correctly, reconciles a second opinion and writes\n   Indic-script words in Latin letters the way people type them (Hinglish,\n   Tanglish …) — never translating them. transliterate: \"native\" keeps the\n   original script. rewrite: false stops after the dictionary.","hierarchy":{"lvl0":"Features","lvl1":"The transcribe inference type","lvl2":"Correction","lvl3":""}},
 {"objectID":"7c55169273332aab63f54e94bae343516bd630e062d60a4d427d004b72ee287e","title":"Nothing is ever dropped","url":"/docs/features/transcribe#nothing-is-ever-dropped","content":"Every layer fails open, and none may lose words:\nA rewrite that drops more than maxDropRatio of the words is discarded and\n  the dictionary-applied text is kept.\nA rewrite that times out, fails or returns nothing keeps the text it was\n  given.\nA guard that times out or fails applies the dictionary as is.\nWhen streaming, a final pass that returns less than the live text keeps the\n  live text, and an utterance no engine could read is reported as a silence\n  event with its length rather than disappearing.\nraw always holds the engine's own text, so a caller can show or diff it.\n\nEach of those outcomes is written to steps, so a degraded path is visible\nrather than silent.","hierarchy":{"lvl0":"Features","lvl1":"The transcribe inference type","lvl2":"Nothing is ever dropped","lvl3":""}},
 {"objectID":"c0207627f5f88b7914c2fbe84a641879ece05ae20f7e2194a9f0bc29c24cfd65","title":"Fallback and second opinion","url":"/docs/features/transcribe#fallback-and-second-opinion","content":"Fallback takes the request (or, when streaming, the utterance) when the\n  primary engine cannot read it: when defaults to [\"unsure\", \"empty\"] — a\n  low language-detection score, or no text — and \"error\" adds engine\n  failures. engine.fallbackUsed says when it ran.\nSecond opinion runs a second engine on the same audio in parallel; the\n  rewrite reconciles the two transcripts, taking each language's words from the\n  engine that reads it better. When the primary reports an unsure language,\n  or no text, the second opinion's text leads. With rewrite: false that is\n  all it does — there is no reconciliation without the rewrite.","hierarchy":{"lvl0":"Features","lvl1":"The transcribe inference type","lvl2":"Fallback and second opinion","lvl3":""}},
 {"objectID":"c903ddf3ec717cab3ff063e2bbb762f634ff09b05e5199c35d95283aba82f2f6","title":"Streaming","url":"/docs/features/transcribe#streaming","content":"Engines with a native stream use it; every other engine is streamed on the\nclient by a chunked adapter (an energy gate finds utterances, the current\nutterance is re-transcribed every intervalMs, and the prefix two passes agree\non is committed). streaming.mode: \"chunked\" forces the adapter.\n\n| Event        | Carries                                                                                        |\n| ------------ | ---------------------------------------------------------------------------------------------- |\n| interim    | The utterance's current text, split into committed (stable) and tail (may still change). |\n| language   | The detected language and whether the engine was sure.                                         |\n| final      | The utterance's final text, segment and words, the engine, and whether the fallback took it.   |\n| correcting | Partial text of the streamed rewrite.                                                          |\n| corrected  | The corrected text, its raw, decisions, steps and timings.                                   |\n| silence    | An utterance no engine could read, with its length.                                            |\n| error      | A message, and recoverable: true when the stream carries on with the next utterance.         |\n\nTuning (streaming): sampleRate (16000), endSilenceMs (750), intervalMs\n(900), softCutSeconds (16), maxUtteranceSeconds (28), onsetRms (0.006),\nprerollMs (350), minSpeechMs (250).","hierarchy":{"lvl0":"Features","lvl1":"The transcribe inference type","lvl2":"Streaming","lvl3":""}},
@@ -5634,13 +5638,13 @@
 {"objectID":"87ba20ffd62f6d99291ed1b32232cafab1ea70cfc060f5a59ea5e374b16de169","title":"Empty content with small maxTokens on gpt-oss-120b","url":"/docs/getting-started/providers/cerebras#empty-content-with-small-maxtokens-on-gpt-oss-120b","content":"gpt-oss-120b is a reasoning model: it spends its first tokens on a\nreasoning channel before emitting content. With a tight budget\n(e.g. maxTokens: 50) the entire budget goes to reasoning,\nfinish_reason is length, and content is empty. Give reasoning\nprompts a few hundred tokens of headroom.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cerebras Provider Guide","lvl2":"Empty content with small maxTokens on gpt-oss-120b","lvl3":""}},
 {"objectID":"12792b087dafedbd6d724cb4cfd2f5365fcea8da18d2d14493ff688b79efda80","title":"404 \"model not found\" for llama/qwen models","url":"/docs/getting-started/providers/cerebras#404-model-not-found-for-llamaqwen-models","content":"Those models are retired, and so is gemma-4-31b. The catalog carries\ngpt-oss-120b only — verify the live roster with an authenticated\nGET https://api.cerebras.ai/v1/models.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cerebras Provider Guide","lvl2":"404 \"model not found\" for llama/qwen models","lvl3":""}},
 {"objectID":"c122483695ccba53489f6d1403f7b78fa2f9d48c9a35220e08425516b5521183","title":"See Also","url":"/docs/getting-started/providers/cerebras#see-also","content":"Groq Provider — sibling speed-focused OpenAI-compat provider\nxAI Grok Provider — sibling OpenAI-compat with Grok 3\nTier 2 catalog entry guide — how this provider is wired internally\n\nNeed Help? Open a GitHub Discussion or issue.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cerebras Provider Guide","lvl2":"See Also","lvl3":""}},
-{"objectID":"23fce3bd9a49c931e827ca107b2e9a56b785e0195bba8b851cc0faba57202684","title":"Charm Hyper Provider Guide","url":"/docs/getting-started/providers/charm-hyper","content":"Charm Hyper Provider Guide\n\nCharm Hyper is a Tier-2 catalog provider: OpenAI-wire-compatible with no\nbehavioural quirks, so its entire integration is one JSON file\n(src/lib/providers/catalog/charm-hyper.json) rather than hand-written code.\nThat file is the source of truth for everything on this page.\n\nVerification status: docs- and roster-verified, not live-verified. This\nentry was onboarded credential-free — everything below comes from Hyper's\nown public documentation and an unauthenticated GET /v1/models call. No\nAPI key was ever created, and no authenticated or POST request was made\nto Hyper — so no chat call, tool call, structured-output call, or error\nresponse was ever actually exercised. Capabilities below reflect what the\ndocs state or omit, not what a live probe proved. Run the\nTier-2 live verification steps\n(roster probe with a real key, billing check, capability probes, live\nmatrix) before relying on this entry in production, and fill in\nevidence.liveMatrix when that happens.\n\nKey Facts\nProvider id: charm-hyper\nProtocol: OpenAI-compatible (/v1/chat/completions). Hyper also\n  documents an Anthropic-compatible /v1/messages endpoint, which this entry\n  does not use.\nBase URL: https://hyper.charm.land/v1\nDefault model: deepseek-v4-pro — this is not a vendor-declared\n  default; Hyper's docs never label any model \"default\" or \"recommended\". It\n  is the model value used in the worked example on Hyper's own Chat\n  Completions reference page. Hyper's homepage quick-start uses a different\n  example model (deepseek-v4-flash) — reconfirm the right default once a\n  key is available.\nModels in catalog: 23 (the full public roster as of 2026-09-28)\nStreaming: documented (stream: true, SSE, terminal [DONE])\nTool calling: documented — tools is a listed request parameter\n  (\"Tool/function definitions\"), but no worked tool-call example exists in\n  the docs and it was never probed live, so toolsWithStreaming is left\n  false pending verification.\nStructured output: not documented at all — the terms\n  response_format, \"JSON mode\" and \"structured output\" never appear\n  anywhere in Hyper's API docs. structuredOutput and\n  structuredOutputWithTools are both false in the catalog.\nEmbeddings: not documented / not supported\nBilling: free-tier — $0/month, 100 Hypercredits/month (1 Hypercredit ≈\n  $0.05); paid subscription and prepaid bundles also exist\nKey format: keys start with sk-hyper-\n\nQuick Start\nGet an API key\nVisit: https://hyper.charm.land/auth?mode=signup and sign up (Google,\n   GitHub, or email)\nCopy an API key from the Hyper Dashboard — it starts with sk-hyper-\nFree plan: $0/month, 100 Hypercredits/month; paid subscription\n   ($20/month, 250 Hypercredits/day) and prepaid bundles (never expire) are\n   also available\nSet CHARM_HYPER_API_KEY in your .env file\nConfigure\nUse it\n\nPer-request credentials work as they do for every provider:\n\nModels\n\nAll 23 entries come from the public, unauthenticated GET\nhttps://hyper.charm.land/v1/models response (2026-09-28) — context window,\nmax output tokens, vision and pricing are the vendor's own served metadata,\nnot estimates.\n\n| Model                    | Context   | Max out | Vision | $/M in · out (cached in)          | Notes                                                                                                                                                                                               |\n| ------------------------ | --------- | ------- | ------ | --------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| deepseek-v4-pro ⭐     | 1M        | 384K    | no     | $2.40 / $4.80 ($0.20)             | Default for this entry — Hyper's own Chat Completions reference example.                                                                                                                            |\n| deepseek-v4.1-flash    | 1,048,576 | 32,768  | yes    | $0.30 / $1.20 ($0.03)             | Newest DeepSeek Flash snapshot; reasoning effort low/high/xhigh.                                                                                                                                    |\n| deepseek-v4-flash      | 1M        | 384K    | no     | $0.20 / $0.40 ($0.04)             | Homepage quick-start's example model. Hyper's marketing page lists this id under \"attachment support\", but the live roster reports vision: false — a documented inconsistency, not resolved here. |\n| deepseek-v4-pro-0813   | 1M        | 262,144 | no     | $1.437216 / $4.311648 ($0.047907) | Dated pin of the Pro line.                                                                                                                                                                          |\n| deepseek-v4-flash-0731 | 1M        | 384K    | no     | $0.44 / $1.32 ($0.044)            | Dated pin of the Flash line.               ","hierarchy":{"lvl0":"Getting Started","lvl1":"Charm Hyper Provider Guide","lvl2":"","lvl3":""}},
+{"objectID":"23fce3bd9a49c931e827ca107b2e9a56b785e0195bba8b851cc0faba57202684","title":"Charm Hyper Provider Guide","url":"/docs/getting-started/providers/charm-hyper","content":"Charm Hyper Provider Guide\n\nCharm Hyper is a Tier-2 catalog provider: OpenAI-wire-compatible with no\nbehavioural quirks, so its entire integration is one JSON file\n(src/lib/providers/catalog/charm-hyper.json) rather than hand-written code.\nThat file is the source of truth for everything on this page.\n\nVerification status: docs- and roster-verified, not live-verified. This\nentry was onboarded credential-free — everything below comes from Hyper's\nown public documentation and an unauthenticated GET /v1/models call. No\nAPI key was ever created, and no authenticated or POST request was made\nto Hyper — so no chat call, tool call, structured-output call, or error\nresponse was ever actually exercised. Capabilities below reflect what the\ndocs state or omit, not what a live probe proved. Run the\nTier-2 live verification steps\n(roster probe with a real key, billing check, capability probes, live\nmatrix) before relying on this entry in production, and fill in\nevidence.liveMatrix when that happens.\n\nKey Facts\nProvider id: charm-hyper\nProtocol: OpenAI-compatible (/v1/chat/completions). Hyper also\n  documents an Anthropic-compatible /v1/messages endpoint, which this entry\n  does not use.\nBase URL: https://hyper.charm.land/v1\nDefault model: deepseek-v4-pro — this is not a vendor-declared\n  default; Hyper's docs never label any model \"default\" or \"recommended\". It\n  is the model value used in the worked example on Hyper's own Chat\n  Completions reference page. Hyper's homepage quick-start uses a different\n  example model (deepseek-v4-flash) — reconfirm the right default once a\n  key is available.\nModels in catalog: 23 (the full public roster as of 2026-09-28)\nStreaming: documented (stream: true, SSE, terminal [DONE])\nTool calling: documented — tools is a listed request parameter\n  (\"Tool/function definitions\"), but no worked tool-call example exists in\n  the docs and it was never probed live, so toolsWithStreaming is left\n  false pending verification.\nStructured output: not documented at all — the terms\n  response_format, \"JSON mode\" and \"structured output\" never appear\n  anywhere in Hyper's API docs. structuredOutput and\n  structuredOutputWithTools are both false in the catalog.\nEmbeddings: not documented / not supported\nBilling: free-tier — $0/month, 100 Hypercredits/month (1 Hypercredit ≈\n  $0.05); paid subscription and prepaid bundles also exist\nKey format: keys start with sk-hyper-\n\nQuick Start\nGet an API key\nVisit: https://hyper.charm.land/auth?mode=signup and sign up (Google,\n   GitHub, or email)\nCopy an API key from the Hyper Dashboard — it starts with sk-hyper-\nFree plan: $0/month, 100 Hypercredits/month; paid subscription\n   ($20/month, 250 Hypercredits/day) and prepaid bundles (never expire) are\n   also available\nSet CHARM_HYPER_API_KEY in your .env file\nConfigure\nUse it\n\nPer-request credentials work as they do for every provider:\n\nModels\n\nAll 23 entries come from the public, unauthenticated GET\nhttps://hyper.charm.land/v1/models response (2026-09-28) — context window,\nmax output tokens, vision and pricing are the vendor's own served metadata,\nnot estimates. The one exception is the max output of four models\n(kimi-k2-thinking, gpt-oss-120b, gemma-4-26b-a4b-it, minimax-m2.7),\nshown as n/a: the roster's figure is a fixed fraction of the context window\n(a tenth for three of them) and reads like a gateway default, so the catalog\nrecords no max output for them and the SDK does not cap maxTokens for those\nmodels.\n\n| Model                    | Context   | Max out | Vision | $/M in · out (cached in)          | Notes                                                                                                                                                                                               |\n| ------------------------ | --------- | ------- | ------ | --------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| deepseek-v4-pro ⭐     | 1M        | 384K    | no     | $2.40 / $4.80 ($0.20)             | Default for this entry — Hyper's own Chat Completions reference example.                                                                                                                            |\n| deepseek-v4.1-flash    | 1,048,576 | 32,768  | yes    | $0.30 / $1.20 ($0.03)             | Newest DeepSeek Flash snapshot; reasoning effort low/high/xhigh.                                                                                                                                    |\n| deepseek-v4-flash      | 1M        | 384K    | no     | $0.20 / $0.40 ($0.04)             | Homepage quick-start's example model. Hyper's marketing page lists this id under \"attachment support\", but the live roster reports vision: false — a documented inconsistency, not resolved here. |\n| deepseek-v4-pro-0813   | 1M        | 262,144 | no     | $1.437216 / $4.","hierarchy":{"lvl0":"Getting Started","lvl1":"Charm Hyper Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"d3aae81e637e5217d4f8060339e76d01d2f5fd4710a1e20516001d1afa4a02be","title":"Charm Hyper Provider Guide","url":"/docs/getting-started/providers/charm-hyper#charm-hyper-provider-guide","content":"Charm Hyper is a Tier-2 catalog provider: OpenAI-wire-compatible with no\nbehavioural quirks, so its entire integration is one JSON file\n(src/lib/providers/catalog/charm-hyper.json) rather than hand-written code.\nThat file is the source of truth for everything on this page.\n\nVerification status: docs- and roster-verified, not live-verified. This\nentry was onboarded credential-free — everything below comes from Hyper's\nown public documentation and an unauthenticated GET /v1/models call. No\nAPI key was ever created, and no authenticated or POST request was made\nto Hyper — so no chat call, tool call, structured-output call, or error\nresponse was ever actually exercised. Capabilities below reflect what the\ndocs state or omit, not what a live probe proved. Run the\nTier-2 live verification steps\n(roster probe with a real key, billing check, capability probes, live\nmatrix) before relying on this entry in production, and fill in\nevidence.liveMatrix when that happens.","hierarchy":{"lvl0":"Getting Started","lvl1":"Charm Hyper Provider Guide","lvl2":"Charm Hyper Provider Guide","lvl3":""}},
 {"objectID":"0669d2bd0baac5391926cfe9f6baee491ae270a7d35bf6e11cd963c840f0727a","title":"Key Facts","url":"/docs/getting-started/providers/charm-hyper#key-facts","content":"Provider id: charm-hyper\nProtocol: OpenAI-compatible (/v1/chat/completions). Hyper also\n  documents an Anthropic-compatible /v1/messages endpoint, which this entry\n  does not use.\nBase URL: https://hyper.charm.land/v1\nDefault model: deepseek-v4-pro — this is not a vendor-declared\n  default; Hyper's docs never label any model \"default\" or \"recommended\". It\n  is the model value used in the worked example on Hyper's own Chat\n  Completions reference page. Hyper's homepage quick-start uses a different\n  example model (deepseek-v4-flash) — reconfirm the right default once a\n  key is available.\nModels in catalog: 23 (the full public roster as of 2026-09-28)\nStreaming: documented (stream: true, SSE, terminal [DONE])\nTool calling: documented — tools is a listed request parameter\n  (\"Tool/function definitions\"), but no worked tool-call example exists in\n  the docs and it was never probed live, so toolsWithStreaming is left\n  false pending verification.\nStructured output: not documented at all — the terms\n  response_format, \"JSON mode\" and \"structured output\" never appear\n  anywhere in Hyper's API docs. structuredOutput and\n  structuredOutputWithTools are both false in the catalog.\nEmbeddings: not documented / not supported\nBilling: free-tier — $0/month, 100 Hypercredits/month (1 Hypercredit ≈\n  $0.05); paid subscription and prepaid bundles also exist\nKey format: keys start with sk-hyper-","hierarchy":{"lvl0":"Getting Started","lvl1":"Charm Hyper Provider Guide","lvl2":"Key Facts","lvl3":""}},
 {"objectID":"510aa5bc0089f7d94fa7ea5b44a48aeb4f4ee41f32be2988d05632f5a64f3f42","title":"1. Get an API key","url":"/docs/getting-started/providers/charm-hyper#1-get-an-api-key","content":"Visit: https://hyper.charm.land/auth?mode=signup and sign up (Google,\n   GitHub, or email)\nCopy an API key from the Hyper Dashboard — it starts with sk-hyper-\nFree plan: $0/month, 100 Hypercredits/month; paid subscription\n   ($20/month, 250 Hypercredits/day) and prepaid bundles (never expire) are\n   also available\nSet CHARM_HYPER_API_KEY in your .env file","hierarchy":{"lvl0":"Getting Started","lvl1":"Charm Hyper Provider Guide","lvl2":"1. Get an API key","lvl3":""}},
 {"objectID":"d2c4f74bf51efdd34688df3eca500f2f8c4d1bb4169a3058381466dc61982842","title":"3. Use it","url":"/docs/getting-started/providers/charm-hyper#3-use-it","content":"`bash","hierarchy":{"lvl0":"Getting Started","lvl1":"Charm Hyper Provider Guide","lvl2":"3. Use it","lvl3":""}},
 {"objectID":"86caff74d39f4f3358d855ac059ac496e308e0d76d2b5a2ff3c50ad43b5b6c90","title":"CLI","url":"/docs/getting-started/providers/charm-hyper#cli","content":"npx @juspay/neurolink generate \"Hello\" --provider charm-hyper\ntypescript\nawait neurolink.generate({\n  input: { text: \"Hello\" },\n  provider: \"charm-hyper\",\n  credentials: { \"charm-hyper\": { apiKey: process.env.CHARMHYPERAPI_KEY } },\n});\n`","hierarchy":{"lvl0":"Getting Started","lvl1":"Charm Hyper Provider Guide","lvl2":"CLI","lvl3":""}},
-{"objectID":"a0302f99aa0bb9839396ece419c976e681649e9eae4939517334dbd917b54060","title":"Models","url":"/docs/getting-started/providers/charm-hyper#models","content":"All 23 entries come from the public, unauthenticated GET\nhttps://hyper.charm.land/v1/models response (2026-09-28) — context window,\nmax output tokens, vision and pricing are the vendor's own served metadata,\nnot estimates.\n\n| Model                    | Context   | Max out | Vision | $/M in · out (cached in)          | Notes                                                                                                                                                                                               |\n| ------------------------ | --------- | ------- | ------ | --------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| deepseek-v4-pro ⭐     | 1M        | 384K    | no     | $2.40 / $4.80 ($0.20)             | Default for this entry — Hyper's own Chat Completions reference example.                                                                                                                            |\n| deepseek-v4.1-flash    | 1,048,576 | 32,768  | yes    | $0.30 / $1.20 ($0.03)             | Newest DeepSeek Flash snapshot; reasoning effort low/high/xhigh.                                                                                                                                    |\n| deepseek-v4-flash      | 1M        | 384K    | no     | $0.20 / $0.40 ($0.04)             | Homepage quick-start's example model. Hyper's marketing page lists this id under \"attachment support\", but the live roster reports vision: false — a documented inconsistency, not resolved here. |\n| deepseek-v4-pro-0813   | 1M        | 262,144 | no     | $1.437216 / $4.311648 ($0.047907) | Dated pin of the Pro line.","hierarchy":{"lvl0":"Getting Started","lvl1":"Charm Hyper Provider Guide","lvl2":"Models","lvl3":""}},
+{"objectID":"a0302f99aa0bb9839396ece419c976e681649e9eae4939517334dbd917b54060","title":"Models","url":"/docs/getting-started/providers/charm-hyper#models","content":"All 23 entries come from the public, unauthenticated GET\nhttps://hyper.charm.land/v1/models response (2026-09-28) — context window,\nmax output tokens, vision and pricing are the vendor's own served metadata,\nnot estimates. The one exception is the max output of four models\n(kimi-k2-thinking, gpt-oss-120b, gemma-4-26b-a4b-it, minimax-m2.7),\nshown as n/a: the roster's figure is a fixed fraction of the context window\n(a tenth for three of them) and reads like a gateway default, so the catalog\nrecords no max output for them and the SDK does not cap maxTokens for those\nmodels.\n\n| Model                    | Context   | Max out | Vision | $/M in · out (cached in)          | Notes                                                                                                                                                                                               |\n| ------------------------ | --------- | ------- | ------ | --------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| deepseek-v4-pro ⭐     | 1M        | 384K    | no     | $2.40 / $4.80 ($0.20)             | Default for this entry — Hyper's own Chat Completions reference example.                                                                                                                            |\n| deepseek-v4.1-flash    | 1,048,576 | 32,768  | yes    | $0.30 / $1.20 ($0.03)             | Newest DeepSeek Flash snapshot; reasoning effort low/high/xhigh.                                                                                                                                    |\n| deepseek-v4-flash      | 1M        | 384K    | no     | $0.20 / $0.40 ($0.04)             | Homepage quick-start's example model.","hierarchy":{"lvl0":"Getting Started","lvl1":"Charm Hyper Provider Guide","lvl2":"Models","lvl3":""}},
 {"objectID":"a55d9b563648ce5a26352a02c15b7226d15f4afb58eec84ad5df54e81c7aa11b","title":"Verification status","url":"/docs/getting-started/providers/charm-hyper#verification-status","content":"Tier-2 onboarding normally requires a live capability sweep, gated by pnpm\nrun verify:provider-onboarding. This entry was deliberately onboarded\nwithout a key, so that step has not run yet:\n\n| Probe                   | Result                                                                                                                                                                                 |\n| ----------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Roster                  | Unauthenticated GET /v1/models, HTTP 200, 2026-09-28 — 23 models, matches Hyper's own docs statement that this endpoint needs no auth (\"Browse the catalog — no auth required\"). |\n| Auth rejection          | Not probed — would require a real or deliberately-invalid API key, i.e. a credential, which this onboarding pass never created.                                                        |\n| Tool calling            | Documented as an accepted parameter (tools, \"Tool/function definitions\") on the Chat Completions reference page; no worked example, no live call.                                    |\n| Structured output       | Not documented anywhere in Hyper's API docs — response_format never appears.                                                                                                     |\n| Streaming + tools combo | Not documented and not probed — toolsWithStreaming is false pending verification.                                                                                                  |\n| Live capability sweep   | Not run. evidence.liveMatrix is null.","hierarchy":{"lvl0":"Getting Started","lvl1":"Charm Hyper Provider Guide","lvl2":"Verification status","lvl3":""}},
 {"objectID":"356875ccab592037d9c5374d90855baafc5b4c8e5cc98acadcc5653f42b25bd4","title":"Troubleshooting","url":"/docs/getting-started/providers/charm-hyper#troubleshooting","content":"| Symptom                                | Cause                                                                        | Fix                                                                                             |\n| -------------------------------------- | ---------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------- |\n| Invalid Charm Hyper API key          | CHARM_HYPER_API_KEY unset, wrong, or missing the sk-hyper- prefix        | Check the key in the Hyper Dashboard at https://hyper.charm.land                                |\n| Model not found                        | The roster changed since 2026-09-28                                          | Pick a current id from the unauthenticated GET /v1/models roster                              |\n| Structured output silently unavailable | structuredOutput is false — Hyper's docs never mention response_format | Use generate() without schema, or verify support live before relying on it                  |\n| Unsure if tool calls stream correctly  | toolsWithStreaming is false — never documented or probed                 | Test with stream: false first, or run the live matrix before depending on streamed tool calls |","hierarchy":{"lvl0":"Getting Started","lvl1":"Charm Hyper Provider Guide","lvl2":"Troubleshooting","lvl3":""}},
 {"objectID":"0eaef58ca3910c8f3b38b4c9615a51d7771d87c84ac285d4109621ca23c43144","title":"See also","url":"/docs/getting-started/providers/charm-hyper#see-also","content":"Provider setup overview\nAll providers\nTier-2 onboarding — how this provider's JSON becomes a working integration, and what live verification still needs to happen\nProvider feature compatibility","hierarchy":{"lvl0":"Getting Started","lvl1":"Charm Hyper Provider Guide","lvl2":"See also","lvl3":""}},
@@ -5655,8 +5659,8 @@
 {"objectID":"5862cb134695d7821bb2656e8d0c7c1ffee1ff3f295c36858f60fa939ba2facd","title":"Verification status","url":"/docs/getting-started/providers/chutes#verification-status","content":"Tier-2 onboarding normally requires a live-key capability sweep\n(docs/provider-integration/tiers/tier-2-catalog-entry.md,\n\"Live verification\"), gated by pnpm run verify:provider-onboarding. This\nentry was added under a credential-free onboarding pass instead: no\nChutes API key was ever created or used, and no authenticated or POST\nrequest was sent to any Chutes endpoint. Everything below is either a public\ndocument or a single unauthenticated GET.\n\n| Probe                     | Result                                                                                                                                                                                                                                                                                     |\n| ------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |\n| Roster                    | Unauthenticated GET https://llm.chutes.ai/v1/models, HTTP 200, 14 models, re-run 2026-09-28. /v1/models is itself documented by Chutes as a \"public catalog (no key required).\"                                                                                                        |\n| Wire shape                | Chutes' own homepage Python sample posts to https://llm.chutes.ai/v1/chat/completions with model + messages + Authorization: Bearer $CHUTES_API_KEY + stream: true — never executed here, only read.                                                                             |\n| Tools / structured output | Confirmed from docs only: the \"Function Calling, Agents, and Tool Use\" guide shows OpenAI-shaped tools/tool_choice/tool_call_id, and the same guide shows a separate response_format: {\"type\": \"json_object\"} JSON-mode example.","hierarchy":{"lvl0":"Getting Started","lvl1":"Chutes Provider Guide","lvl2":"Verification status","lvl3":""}},
 {"objectID":"34c0be7e44128ded319ab20090fe5bfecd98725996a4350660f6dad0932a5601","title":"Troubleshooting","url":"/docs/getting-started/providers/chutes#troubleshooting","content":"| Symptom                                                    | Cause                                                                                            | Fix                                                                                                                                                         |\n| ---------------------------------------------------------- | ------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Requests hit an anonymous rate limit (HTTP 429)            | No token, or an invalid token, was sent                                                          | Set CHUTES_API_KEY, and send it as Authorization: Bearer <key> — X-API-Key is documented as unsupported for inference calls                           |\n| Model not found                                            | The roster changed since 2026-09-28                                                              | Pick a current id from the unauthenticated GET /v1/models roster                                                                                          |\n| Structured output silently dropped when tools are used     | structuredOutputWithTools is false — no combined probe was possible                          | This is intentional pending a live-key check, not a bug; NeuroLink's runtime conflict retry drops structured output and retries rather than losing the turn |\n| Unsure whether a specific model supports tools / JSON mode | capabilities.tools is \"model-dependent\" — two roster ids ship no supported_features at all | Check the model's entry in GET /v1/models, or consult the Models table above                                                                   |","hierarchy":{"lvl0":"Getting Started","lvl1":"Chutes Provider Guide","lvl2":"Troubleshooting","lvl3":""}},
 {"objectID":"f743dd0fbfd3517546cecc2240a8513ad5a47938ceae45eec2ad4fe45aa06d9f","title":"See also","url":"/docs/getting-started/providers/chutes#see-also","content":"Provider setup overview\nAll providers\nTier-2 onboarding — how this provider's JSON becomes a working integration, and what live verification still needs to happen\nProvider feature compatibility","hierarchy":{"lvl0":"Getting Started","lvl1":"Chutes Provider Guide","lvl2":"See also","lvl3":""}},
-{"objectID":"1fb3d0f4a1510905b14436d78a4928a8e825b3dbdb324b6cb7c9a340e088d0f6","title":"Cloudflare Clef Provider Guide","url":"/docs/getting-started/providers/cloudflare-clef","content":"Cloudflare Clef Provider Guide\n\nA provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR,\nPerplexity and OpenAI Decisions,\nfrom Cloudflare's Clef models on Workers AI, and it also reads images. It emits\nno text at all.\n\nThis is not the Cloudflare Workers AI text provider\n(cloudflare, which runs chat models). The two are separate providers that\nread the same CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID; see\nOne token, two providers.\n\nRead Limits before you rely on it. Workers AI's state window\nchanged in October 2026 (hosted service or model: unknown). Until 2026-10-04 it\nread only about the first 2,048 tokens of the state, far less than the 64K\ncontext Cloudflare documents, and ignored the rest without an error; on\n2026-10-07 it read states of at least 190,153 tokens (clef-flash) and 52,154\n(clef). NeuroLink refuses a state it estimates at more than 16,000 tokens,\ninside what was read in full, and a longer wait than a fail-open feature can\nafford.\n\nOverview\n\nClef is a decision model: it reads a state and a set of typed questions and\nreturns a probability for every allowed answer, with no free-form text and no\nreasoning tokens to wait for. Cloudflare publishes two sizes, clef (27B) and\nclef-flash (9B), and says it is releasing their weights under the Apache 2.0\nlicense. Both answered in about a second from a developer machine. Cloudflare's\npages call them available, and do not say whether they are generally available\nor in beta.\n\nKey Facts\n\n|                         |                                                                                                                                                                                       |\n| ----------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Provider id             | cloudflare-clef                                                                                                                                                                     |\n| Models                  | clef (27B, the default) and clef-flash (9B). @cf/cloudflare/clef and @cf/cloudflare/clef-flash are accepted too                                                               |\n| Question types          | boolean (sent as noul), choice (2 to 255 options), score (2 to 10 levels)                                                                                                     |\n| Questions per request   | 64                                                                                                                                                                                    |\n| Images                  | up to 4 per request: PNG, JPEG or WebP. No video                                                                                                                                      |\n| State                   | read in full up to at least 190,153 tokens (clef-flash) and 52,154 (clef) on 2026-10-07; about 2,048 until 2026-10-04. NeuroLink refuses a state it estimates at more than 16,000 |\n| Request size            | 256,000 bytes in NeuroLink, base64 image data included                                                                                                                                |\n| Endpoint                | https://api.cloudflare.com/client/v4/accounts/<account id>/ai/run/@cf/cloudflare/<model>                                                                                            |\n| Credentials             | an API token with Workers AI permission, and the account id                                                                                                                           |\n| Price                   | $0.24 per million input tokens for clef, $0.09 for clef-flash; no output price is listed (Cloudflare's Workers AI pricing page)                                                   |\n| Measured latency        | 2026-10-03: 0.3 to 1.0 s for a small request; 64 questions: 1.1 s (clef-flash), 1.3 s (clef); 2026-10-04: 1.5 s / 2.3 s                                                           |\n| Default decide provider | second to last: after TypeSafe, Laya, XOR, SubQ and Perplexity, before OpenAI Decisions                                                                                               |\n| Verified live           | 2026-10-03 and 2026-10-04, with a real account (see Limits)                                                                                                                |\n\nQuick Start\nGet a token and your account id\n\nCreate an API token with the Workers AI: Read + Write permission at\ndash.cloudflare.com/profile/api-tokens, and copy your account id from\nthe dashboard URL or the \"Account ID\" panel.\nConfigure\n\nThe base URL must be https://; http:// is accepted only for localhost,\n127.0.0.1 and [::1], because every request ca","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"","lvl3":""}},
-{"objectID":"eebb9363ccd61130d0be9464df57f02fd61d00b6ff81821dd3162c030c0bc0db","title":"Cloudflare Clef Provider Guide","url":"/docs/getting-started/providers/cloudflare-clef#cloudflare-clef-provider-guide","content":"A provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR,\nPerplexity and OpenAI Decisions,\nfrom Cloudflare's Clef models on Workers AI, and it also reads images. It emits\nno text at all.\n\nThis is not the Cloudflare Workers AI text provider\n(cloudflare, which runs chat models). The two are separate providers that\nread the same CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID; see\nOne token, two providers.\n\nRead Limits before you rely on it. Workers AI's state window\nchanged in October 2026 (hosted service or model: unknown). Until 2026-10-04 it\nread only about the first 2,048 tokens of the state, far less than the 64K\ncontext Cloudflare documents, and ignored the rest without an error; on\n2026-10-07 it read states of at least 190,153 tokens (clef-flash) and 52,154\n(clef). NeuroLink refuses a state it estimates at more than 16,000 tokens,\ninside what was read in full, and a longer wait than a fail-open feature can\nafford.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"Cloudflare Clef Provider Guide","lvl3":""}},
+{"objectID":"1fb3d0f4a1510905b14436d78a4928a8e825b3dbdb324b6cb7c9a340e088d0f6","title":"Cloudflare Clef Provider Guide","url":"/docs/getting-started/providers/cloudflare-clef","content":"Cloudflare Clef Provider Guide\n\nA provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR,\nSubQ, Perplexity,\nOpenAI Decisions and ScaleDown,\nfrom Cloudflare's Clef models on Workers AI, and it also reads images. It emits\nno text at all.\n\nThis is not the Cloudflare Workers AI text provider\n(cloudflare, which runs chat models). The two are separate providers that\nread the same CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID; see\nOne token, two providers.\n\nRead Limits before you rely on it. Workers AI's state window\nchanged in October 2026 (hosted service or model: unknown). Until 2026-10-04 it\nread only about the first 2,048 tokens of the state, far less than the 64K\ncontext Cloudflare documents, and ignored the rest without an error; on\n2026-10-07 it read states of at least 190,153 tokens (clef-flash) and 52,154\n(clef). NeuroLink refuses a state it estimates at more than 16,000 tokens,\ninside what was read in full, and a longer wait than a fail-open feature can\nafford.\n\nOverview\n\nClef is a decision model: it reads a state and a set of typed questions and\nreturns a probability for every allowed answer, with no free-form text and no\nreasoning tokens to wait for. Cloudflare publishes two sizes, clef (27B) and\nclef-flash (9B), and says it is releasing their weights under the Apache 2.0\nlicense. Both answered in about a second from a developer machine. Cloudflare's\npages call them available, and do not say whether they are generally available\nor in beta.\n\nKey Facts\n\n|                         |                                                                                                                                                                                       |\n| ----------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Provider id             | cloudflare-clef                                                                                                                                                                     |\n| Models                  | clef (27B, the default) and clef-flash (9B). @cf/cloudflare/clef and @cf/cloudflare/clef-flash are accepted too                                                               |\n| Question types          | boolean (sent as noul), choice (2 to 255 options), score (2 to 10 levels)                                                                                                     |\n| Questions per request   | 64                                                                                                                                                                                    |\n| Images                  | up to 4 per request: PNG, JPEG or WebP. No video                                                                                                                                      |\n| State                   | read in full up to at least 190,153 tokens (clef-flash) and 52,154 (clef) on 2026-10-07; about 2,048 until 2026-10-04. NeuroLink refuses a state it estimates at more than 16,000 |\n| Request size            | 256,000 bytes in NeuroLink, base64 image data included                                                                                                                                |\n| Endpoint                | https://api.cloudflare.com/client/v4/accounts/<account id>/ai/run/@cf/cloudflare/<model>                                                                                            |\n| Credentials             | an API token with Workers AI permission, and the account id                                                                                                                           |\n| Price                   | $0.24 per million input tokens for clef, $0.09 for clef-flash; no output price is listed (Cloudflare's Workers AI pricing page)                                                   |\n| Measured latency        | 2026-10-03: 0.3 to 1.0 s for a small request; 64 questions: 1.1 s (clef-flash), 1.3 s (clef); 2026-10-04: 1.5 s / 2.3 s                                                           |\n| Default decide provider | sixth of eight: after TypeSafe, Laya, XOR, SubQ and Perplexity, before OpenAI Decisions and ScaleDown                                                                                 |\n| Verified live           | 2026-10-03 and 2026-10-04, with a real account (see Limits)                                                                                                                |\n\nQuick Start\nGet a token and your account id\n\nCreate an API token with the Workers AI: Read + Write permission at\ndash.cloudflare.com/profile/api-tokens, and copy your account id from\nthe dashboard URL or the \"Account ID\" panel.\nConfigure\n\nThe base URL must be https://; http:// is accepted only for localhost,\n127.0.0.1 and [::1], because","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"","lvl3":""}},
+{"objectID":"eebb9363ccd61130d0be9464df57f02fd61d00b6ff81821dd3162c030c0bc0db","title":"Cloudflare Clef Provider Guide","url":"/docs/getting-started/providers/cloudflare-clef#cloudflare-clef-provider-guide","content":"A provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR,\nSubQ, Perplexity,\nOpenAI Decisions and ScaleDown,\nfrom Cloudflare's Clef models on Workers AI, and it also reads images. It emits\nno text at all.\n\nThis is not the Cloudflare Workers AI text provider\n(cloudflare, which runs chat models). The two are separate providers that\nread the same CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID; see\nOne token, two providers.\n\nRead Limits before you rely on it. Workers AI's state window\nchanged in October 2026 (hosted service or model: unknown). Until 2026-10-04 it\nread only about the first 2,048 tokens of the state, far less than the 64K\ncontext Cloudflare documents, and ignored the rest without an error; on\n2026-10-07 it read states of at least 190,153 tokens (clef-flash) and 52,154\n(clef). NeuroLink refuses a state it estimates at more than 16,000 tokens,\ninside what was read in full, and a longer wait than a fail-open feature can\nafford.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"Cloudflare Clef Provider Guide","lvl3":""}},
 {"objectID":"d3aa34515e7690cc133be051b57b7fda1e8fdd225ef0dd7a162d91ee2ffa91d8","title":"Overview","url":"/docs/getting-started/providers/cloudflare-clef#overview","content":"Clef is a decision model: it reads a state and a set of typed questions and\nreturns a probability for every allowed answer, with no free-form text and no\nreasoning tokens to wait for. Cloudflare publishes two sizes, clef (27B) and\nclef-flash (9B), and says it is releasing their weights under the Apache 2.0\nlicense. Both answered in about a second from a developer machine. Cloudflare's\npages call them available, and do not say whether they are generally available\nor in beta.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"Overview","lvl3":""}},
 {"objectID":"babb1678dd2f0f63c1db8cf3e5216f3ac6b22a00331fe49058519957948ca5fb","title":"Key Facts","url":"/docs/getting-started/providers/cloudflare-clef#key-facts","content":"|                         |                                                                                                                                                                                       |\n| ----------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Provider id             | cloudflare-clef                                                                                                                                                                     |\n| Models                  | clef (27B, the default) and clef-flash (9B). @cf/cloudflare/clef and @cf/cloudflare/clef-flash are accepted too                                                               |\n| Question types          | boolean (sent as noul), choice (2 to 255 options), score (2 to 10 levels)                                                                                                     |\n| Questions per request   | 64                                                                                                                                                                                    |\n| Images                  | up to 4 per request: PNG, JPEG or WebP. No video                                                                                                                                      |\n| State                   | read in full up to at least 190,153 tokens (clef-flash) and 52,154 (clef) on 2026-10-07; about 2,048 until 2026-10-04.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"Key Facts","lvl3":""}},
 {"objectID":"5f8fca781ea20214c667a17eb2e49fc41c5a9187b58b202f67758f2bfc581c4e","title":"1. Get a token and your account id","url":"/docs/getting-started/providers/cloudflare-clef#1-get-a-token-and-your-account-id","content":"Create an API token with the Workers AI: Read + Write permission at\ndash.cloudflare.com/profile/api-tokens, and copy your account id from\nthe dashboard URL or the \"Account ID\" panel.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"1. Get a token and your account id","lvl3":""}},
@@ -5664,7 +5668,7 @@
 {"objectID":"de4ffe6ae22db4a0719caae8e38b4a0a8a438224c94701a762c9e1f4cc209393","title":"Optional","url":"/docs/getting-started/providers/cloudflare-clef#optional","content":"typescript\n\nconst neurolink = new NeuroLink({\n  credentials: {\n    cloudflareClef: { apiKey: \"…\", accountId: \"…\" }, // baseURL is optional\n  },\n});\n`\n\nBoth the token and the account id are needed: the account id is part of the\nroute, so a token alone does not count as configured.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"Optional","lvl3":""}},
 {"objectID":"0b26521b87bd72415fe58482c8628d4c3d04a49575ca2e89a55f29cf65834ccc","title":"3. Use it","url":"/docs/getting-started/providers/cloudflare-clef#3-use-it","content":"The figures in the comments are what a live call to clef returned for this\nexample on 2026-10-03, rounded; clef-flash answered the same example with\n0.96, 0.82 and 0.50, 2.72. Pick the faster model for one call with\nmodel: \"clef-flash\".","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"3. Use it","lvl3":""}},
 {"objectID":"251d84b27d167ed1d24cad4c91fb45246bf057457ea226ad977f4e581af03d28","title":"Images","url":"/docs/getting-started/providers/cloudflare-clef#images","content":"Pass up to four images beside the state, as a Buffer, a file path or a\ndata:image/…;base64, URL; an http(s) URL is refused. PNG, JPEG and WebP are\nread; image/jpg is accepted as a JPEG alias and sent as image/jpeg. Other formats are refused before they are sent.\nThey travel in their own images array, not inside the state. Cloudflare\n  places them before the state.\nThe Workers AI API accepts no video. Its model page says it does, but the API refuses\n  a video and has no field for one, so NeuroLink refuses a video too.\nPixels: Cloudflare documents 16 megapixels an image. 16.00 megapixels were\n  accepted (4096 × 3900 at 15.97 MP, 8000 × 2000 and 16000 × 1000 at 16.00 MP);\n  16.38 megapixels (4096 × 4000) were refused with a 422. The limit is on the\n  pixel count, not on either side.\nCost: an image costs at most about 1,000 input tokens however large it is.\n  A test request with one 512 × 512 image counted 389 input tokens in all, one\n  with a 1,024 × 1,024 image counted 1,157, and larger images up to the pixel\n  limit counted between 1,125 and 1,157.\nHow many: exactly four tiny PNG/JPEG/WebP images succeeded on both models\n  on 2026-10-04; a fifth was refused with a 422. Four is also NeuroLink's limit.\nNeuroLink's size limit is 256,000 bytes for the encoded request body,\n  including base64 image data. A 195 KB PNG encodes to about 260,000 characters\n  and is refused locally. Resize or compress a photo to well under 190 KB first.\n  The 2026-10-03 API probe accepted a 195 KB PNG and refused a 202 KB one.\n  That boundary has since moved: on 2026-10-05 both models accepted PNG\n  requests with encoded bodies up to 492,485 bytes and refused 532,505 bytes\n  (clef-flash) and 532,499 (clef) (see below), so the 195/202 KB figures are historical and NeuroLink's cap is\n  the stricter limit.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"Images","lvl3":""}},
-{"objectID":"f98057fa646a175aee28fef3b525786c83a66b8ac8489feec28fd872866cf4ab","title":"When NeuroLink uses it","url":"/docs/getting-started/providers/cloudflare-clef#when-neurolink-uses-it","content":"Built-in features that call decide() pick the first decision provider that is\nconfigured, in this order: TypeSafe, Laya, XOR, SubQ, Perplexity, Clef, then\nOpenAI Decisions. A host with none of the first five, but\nwith CLOUDFLARE_API_KEY and\nCLOUDFLARE_ACCOUNT_ID set for the Workers AI text provider, therefore has Clef\nas its default decision provider. Naming provider: \"cloudflare-clef\" reaches it\nwhatever else is configured.\n\nBecause of the state limit below, a built-in feature whose state is longer than\nabout 16,000 estimated tokens (about 60,000 characters of prose) gets a refusal\nfrom Clef, which each of them treats as \"carry on as before\". Clef suits short decisions: routing a request, a\nguardrail on one action, classifying one message or one picture.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"When NeuroLink uses it","lvl3":""}},
+{"objectID":"f98057fa646a175aee28fef3b525786c83a66b8ac8489feec28fd872866cf4ab","title":"When NeuroLink uses it","url":"/docs/getting-started/providers/cloudflare-clef#when-neurolink-uses-it","content":"Built-in features that call decide() pick the first decision provider that is\nconfigured, in this order: TypeSafe, Laya, XOR, SubQ, Perplexity, Clef,\nOpenAI Decisions, then ScaleDown. A host with none of the first five, but\nwith CLOUDFLARE_API_KEY and\nCLOUDFLARE_ACCOUNT_ID set for the Workers AI text provider, therefore has Clef\nas its default decision provider. Naming provider: \"cloudflare-clef\" reaches it\nwhatever else is configured.\n\nBecause of the state limit below, a built-in feature whose state is longer than\nabout 16,000 estimated tokens (about 60,000 characters of prose) gets a refusal\nfrom Clef, which each of them treats as \"carry on as before\". Clef suits short decisions: routing a request, a\nguardrail on one action, classifying one message or one picture.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"When NeuroLink uses it","lvl3":""}},
 {"objectID":"69f3db3e89a8fd474930cba6df7b01125fa2d9dd41aaadf8d8838423632ace30","title":"What is sent to Cloudflare","url":"/docs/getting-started/providers/cloudflare-clef#what-is-sent-to-cloudflare","content":"boolean is the SDK's name for the wire's noul. model is sent although the\npath already names it: Cloudflare's schema marks it required, and a body whose\nmodel differs from the path is refused. A request with no model was accepted\nlive and answered by clef-flash, so sending it matters only for following the\nschema. Cloudflare wraps every answer in its own\nenvelope, and NeuroLink reads the answers out of it:\n\nThe request id is the cf-ai-req-id response header; where that is absent (a\nrejected token never reaches the model) it is the edge's cf-ray.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"What is sent to Cloudflare","lvl3":""}},
 {"objectID":"37eb6d126d3af6ca22085fbf94025a8f23da15691c8f6070b946f5aaa27e2f10","title":"One token, two providers","url":"/docs/getting-started/providers/cloudflare-clef#one-token-two-providers","content":"cloudflare-clef and cloudflare read the same two variables, but each has its\nown credentials slice, so the two cannot be mixed up:\n\n|                 | cloudflare (text)         | cloudflare-clef (decide)   |\n| --------------- | --------------------------- | ---------------------------- |\n| Token           | CLOUDFLARE_API_KEY        | CLOUDFLARE_API_KEY         |\n| Account id      | CLOUDFLARE_ACCOUNT_ID     | CLOUDFLARE_ACCOUNT_ID      |\n| SDK credentials | credentials.cloudflare    | credentials.cloudflareClef |\n| Serves          | generate() and stream() | decide() only              |\n\ncredentials.cloudflare does not configure decide. Setting the two\nvariables for the text provider also lets decide() use Clef when no other\ndecision provider is configured, and no switch turns that off while the two\nvariables are set in the environment. A host that wants the Workers AI text\nprovider without Clef can give the text provider its token and account id only\nthrough credentials.cloudflare, and leave the environment variables unset.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"One token, two providers","lvl3":""}},
 {"objectID":"fee62c91a128411e33fcad60cf048405bd86ff0f4015189d992b70d34fe1b86f","title":"Measured on a real account, October 2026","url":"/docs/getting-started/providers/cloudflare-clef#measured-on-a-real-account-october-2026","content":"The Workers AI endpoint ignored text past about 2,048 tokens until 2026-10-04,\nfar below Cloudflare's documented 64K context, and read far more on 2026-10-07\n(hosted service or model: unknown). Request-size refusals also changed between two measurement dates.\nThe original measurements were on 2026-10-03; the bounded follow-up campaign\nused 133 probe calls on 2026-10-04, with one request at a time.\n\nWhich model: on both clef-flash and clef, the follow-up still read the\nfact at the 2026-10-03 clef-flash lower bound for logs, number lists, digit\narrays and compact JSON, and did not read it 2.5% to 3.1% further on. English prose and\nrandom CJK had already matched on both models. It also checked the question\ncount, pixel and four/five-image boundaries on both models, the id, option,\nscore-level and PNG/JPEG/WebP boundaries on clef, and the image/jpg case on\nboth models. Natural-script\nprose was measured on clef; a reworked many-key object probe matched on both.\nTypeScript, minified JSON and synthetic Devanagari/emoji cuts remain 9B-only\nmeasurements. Historical latency, billing and burst figures retain their dates.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"Measured on a real account, October 2026","lvl3":""}},
@@ -5674,8 +5678,8 @@
 {"objectID":"4214184c06ad6cee88e38838f55c38c5a554a78434494b5d765a1c120eb0e339","title":"Latency and the timeout","url":"/docs/getting-started/providers/cloudflare-clef#latency-and-the-timeout","content":"On 2026-10-03, a small request answered in 0.3 to 1.0 s from a developer\nmachine; 64 questions took 1.1 s on clef-flash and 1.3 s on clef; the\nslowest of 60 requests sent at once took 2.2 s. On 2026-10-04, a\n64-question request with the same questions and a shorter state took 1.5 s on clef-flash and 2.3 s on clef. The default timeout is 5 s plus 250 ms for each full 1,000 estimated state tokens\n(5 s for a short state, 9 s at the 16,000-token limit); on 2026-10-07 large states took about 0.2 s per\n1,000 input tokens (1.7 to 1.9 s for 20,954 tokens on clef-flash; 11.8 s for 52,154 on clef). Pass\ntimeoutMs to change it for one call.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"Latency and the timeout","lvl3":""}},
 {"objectID":"c2d0b99e3c2760e14d8be1f81ba92c025ad96523cdeaa34dd002a748ae9658de","title":"Errors","url":"/docs/getting-started/providers/cloudflare-clef#errors","content":"Cloudflare answers with its own envelope, { \"success\": false, \"errors\": [{ \"code\",\n\"message\" }] }, and a refusal from the model nests a second envelope, with a\ntrailing request id, inside message. NeuroLink flattens both to one line and\ntakes the request id out of the text.\n\n| Status | Code  | Seen for                                                                                  | Kind                                | Retried                        |\n| ------ | ----- | ----------------------------------------------------------------------------------------- | ----------------------------------- | ------------------------------ |\n| 401    | 10000 | a token Cloudflare does not know                                                          | authentication                    | no; trips the breaker          |\n| 403    |       | never seen; a token without Workers AI permission may draw it                             | invalid_request                   | no; does not trip the breaker  |\n| 400    | 5006  | no questions; an unknown question type                                                    | invalid_request                   | no                             |\n| 400    | 7000  | a model path that does not exist (\"No route for that URI\")                                | invalid_request                   | no                             |\n| 400    | 6003  | a body that is not JSON                                                                   | invalid_request                   | no                             |\n| 422    | 5012  | a validation failure: 65 questions, a fifth image, an unknown field, an empty instruction | invalid_request                   | no                             |\n| 413    | 5021  | a request past the estimate above                                                         | max_tokens_exceeded               | no                             |\n| 429    | 3040  | \"Capacity temporarily exceeded\"; no Retry-After","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"Errors","lvl3":""}},
 {"objectID":"ce05a3d9d2ab15b8867c8e721fbfa0f8e1248520e2413765fa7d70bd835e3f85","title":"Troubleshooting","url":"/docs/getting-started/providers/cloudflare-clef#troubleshooting","content":"\"requires the account id\" — set CLOUDFLARE_ACCOUNT_ID or\n  credentials.cloudflareClef.accountId.\n\"The state is ~N tokens; Cloudflare's … model reads at most 1500\" — shorten\n  the state, or send only the part the question is about. Cloudflare would have\n  cut the rest without telling you.\n\"No route for that URI\" — the model name or the base URL is wrong. Clef is\n  served as clef and clef-flash, and the base URL must end in /client/v4.\n\"The Cloudflare base URL must use https:// unless it points at this machine\" —\n  the override is an http:// URL whose host is not localhost, 127.0.0.1 or\n  [::1]. A plain-http request would send the token unencrypted. Use https://,\n  or reach the server through a port-forward to localhost.\n\"Cloudflare answered with a redirect\" — the endpoint replied 301, 302, 307 or 308. The provider does not follow it and does not retry. Set the base URL to the\n  address the redirect leads to.\nA 429 \"Capacity temporarily exceeded\" — retried for you; if it keeps\n  happening, send fewer requests at once.\nA 429 \"you have used up your daily free allocation of 10,000 neurons\" — the\n  Free plan's allowance for the day is gone. It is not retried, because nothing\n  changes until the allowance returns (Cloudflare documents a reset at 00:00 UTC; on\n  this account it came back later, see \"What is not verified\"); upgrade to Workers Paid or wait.\n  Built-in features that use Clef as their default fall back as on any failure.\nAn image is refused as \"The request is N bytes; Cloudflare accepts at most\n  256000\" — Cloudflare counts the base64 text of the image, so NeuroLink\n  refuses it locally under the retained conservative cap. Resize or compress it\n  to well under 190 KB; a 195 KB PNG, which Cloudflare accepted, is about 260,000\n  characters in base64 and is refused here.\ndecide() uses Clef although you never chose it — you have the Workers AI\n  token and account id set and no other decision provider; see\n  When NeuroLink uses it.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"Troubleshooting","lvl3":""}},
-{"objectID":"64d2af754e23d6f77f9ba5991174048f9dc4e8434d0f9a633527785feff8158e","title":"See also","url":"/docs/getting-started/providers/cloudflare-clef#see-also","content":"The decide inference type\nTypeSafe (Jev) Provider Guide\nLaya Provider Guide\nXOR Provider Guide\nPerplexity Decisions Provider Guide\nSubQ Provider Guide\nOpenAI Decisions Provider Guide\nCloudflare Workers AI text provider\nClef on Workers AI\nIntroducing Clef: our open-source decision models, and new RL fine-tuning platform (Cloudflare Blog)","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"See also","lvl3":""}},
-{"objectID":"3803b4ecde578f946b8a4fe5139e3ee640d53281541c50578e8a5d47847d5157","title":"Cloudflare Workers AI Provider Guide","url":"/docs/getting-started/providers/cloudflare","content":"Cloudflare Workers AI Provider Guide\n\nOpen-model inference at the edge via Cloudflare Workers AI\n\nOverview\n\nCloudflare Workers AI\nserves Meta Llama, Mistral, and other open models from Cloudflare's\nglobal GPU cluster. NeuroLink talks to the OpenAI-compatible endpoint.\n\nKey Facts\nProtocol: OpenAI-compatible (/v1/chat/completions)\nDefault base URL: https://api.cloudflare.com/client/v4/accounts/{accountId}/ai/v1\nDefault model: @cf/meta/llama-3.3-70b-instruct-fp8-fast\nStreaming: Yes\nTool calling: Limited (model-dependent)\n\nQuick Start\nGet Credentials\n\nYou need both:\nA Cloudflare Account ID (Cloudflare dashboard → right sidebar)\nA Workers AI API token with the Workers AI Read & Write\n  permission (Profile → API Tokens → Create Token)\nConfigure\nGenerate\n\nSupported Models (sample)\n\n| Model ID                                   | Notes          |\n| ------------------------------------------ | -------------- |\n| @cf/meta/llama-3.3-70b-instruct-fp8-fast | Default        |\n| @cf/meta/llama-3.1-70b-instruct          | Llama 3.1 70B  |\n| @cf/meta/llama-3.1-8b-instruct           | Fast tier      |\n| @cf/meta/llama-3.2-11b-vision-instruct   | Vision-capable |\n\nBrowse: https://developers.cloudflare.com/workers-ai/models\n\nCLI Usage\n\nProvider Aliases\n\n| Alias        | Example                 |\n| ------------ | ----------------------- |\n| cloudflare | --provider cloudflare |\n| cf         | --provider cf         |\n\nConfiguration Reference\n\n| Environment Variable    | Required | Default                                    |\n| ----------------------- | -------- | ------------------------------------------ |\n| CLOUDFLARE_ACCOUNT_ID | Yes      | —                                          |\n| CLOUDFLARE_API_KEY    | Yes      | —                                          |\n| CLOUDFLARE_MODEL      | No       | @cf/meta/llama-3.3-70b-instruct-fp8-fast |\n\nAlso configures decisions\n\nCLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID also configure a separate\nprovider, the Cloudflare Clef provider (cloudflare-clef),\nwhich serves decide() and emits no text. It is the second-to-last decision\nprovider NeuroLink falls back to: built-in features that call decide() use the\nfirst one configured, in the order TypeSafe, Laya, XOR, SubQ, Perplexity, Clef, then\nOpenAI Decisions. So a host\nthat set these two variables only for this text provider, and has no other\ndecision provider configured, has Clef as its default decision provider, and no\nswitch turns that off while they are set.\n\ncredentials.cloudflare does not configure decide; Clef reads its own\nslice, credentials.cloudflareClef. The text provider itself is unchanged: it\nstill serves generate() and stream() with the same variables, models and base\nURL. See\nOne token, two providers.\n\nSee Also\nCloudflare Clef Provider — typed decide() judgments on the same token and account id\nTogether AI Provider\nFireworks Provider","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Workers AI Provider Guide","lvl2":"","lvl3":""}},
+{"objectID":"64d2af754e23d6f77f9ba5991174048f9dc4e8434d0f9a633527785feff8158e","title":"See also","url":"/docs/getting-started/providers/cloudflare-clef#see-also","content":"The decide inference type\nTypeSafe (Jev) Provider Guide\nLaya Provider Guide\nXOR Provider Guide\nPerplexity Decisions Provider Guide\nSubQ Provider Guide\nOpenAI Decisions Provider Guide\nScaleDown Provider Guide\nCloudflare Workers AI text provider\nClef on Workers AI\nIntroducing Clef: our open-source decision models, and new RL fine-tuning platform (Cloudflare Blog)","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Clef Provider Guide","lvl2":"See also","lvl3":""}},
+{"objectID":"3803b4ecde578f946b8a4fe5139e3ee640d53281541c50578e8a5d47847d5157","title":"Cloudflare Workers AI Provider Guide","url":"/docs/getting-started/providers/cloudflare","content":"Cloudflare Workers AI Provider Guide\n\nOpen-model inference at the edge via Cloudflare Workers AI\n\nOverview\n\nCloudflare Workers AI\nserves Meta Llama, Mistral, and other open models from Cloudflare's\nglobal GPU cluster. NeuroLink talks to the OpenAI-compatible endpoint.\n\nKey Facts\nProtocol: OpenAI-compatible (/v1/chat/completions)\nDefault base URL: https://api.cloudflare.com/client/v4/accounts/{accountId}/ai/v1\nDefault model: @cf/meta/llama-3.3-70b-instruct-fp8-fast\nStreaming: Yes\nTool calling: Limited (model-dependent)\n\nQuick Start\nGet Credentials\n\nYou need both:\nA Cloudflare Account ID (Cloudflare dashboard → right sidebar)\nA Workers AI API token with the Workers AI Read & Write\n  permission (Profile → API Tokens → Create Token)\nConfigure\nGenerate\n\nSupported Models (sample)\n\n| Model ID                                   | Notes          |\n| ------------------------------------------ | -------------- |\n| @cf/meta/llama-3.3-70b-instruct-fp8-fast | Default        |\n| @cf/meta/llama-3.1-70b-instruct          | Llama 3.1 70B  |\n| @cf/meta/llama-3.1-8b-instruct           | Fast tier      |\n| @cf/meta/llama-3.2-11b-vision-instruct   | Vision-capable |\n\nBrowse: https://developers.cloudflare.com/workers-ai/models\n\nCLI Usage\n\nProvider Aliases\n\n| Alias        | Example                 |\n| ------------ | ----------------------- |\n| cloudflare | --provider cloudflare |\n| cf         | --provider cf         |\n\nConfiguration Reference\n\n| Environment Variable    | Required | Default                                    |\n| ----------------------- | -------- | ------------------------------------------ |\n| CLOUDFLARE_ACCOUNT_ID | Yes      | —                                          |\n| CLOUDFLARE_API_KEY    | Yes      | —                                          |\n| CLOUDFLARE_MODEL      | No       | @cf/meta/llama-3.3-70b-instruct-fp8-fast |\n\nAlso configures decisions\n\nCLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID also configure a separate\nprovider, the Cloudflare Clef provider (cloudflare-clef),\nwhich serves decide() and emits no text. It is the sixth of the eight decision\nproviders NeuroLink falls back to: built-in features that call decide() use the\nfirst one configured, in the order TypeSafe, Laya, XOR, SubQ, Perplexity, Clef,\nOpenAI Decisions, then ScaleDown. So a host\nthat set these two variables only for this text provider, and has no other\ndecision provider configured, has Clef as its default decision provider, and no\nswitch turns that off while they are set.\n\ncredentials.cloudflare does not configure decide; Clef reads its own\nslice, credentials.cloudflareClef. The text provider itself is unchanged: it\nstill serves generate() and stream() with the same variables, models and base\nURL. See\nOne token, two providers.\n\nSee Also\nCloudflare Clef Provider — typed decide() judgments on the same token and account id\nTogether AI Provider\nFireworks Provider","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Workers AI Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"aafa1a6f2adf2bacc5661cf9ccb309d1e32e300645f037c35a98217b31eb15df","title":"Cloudflare Workers AI Provider Guide","url":"/docs/getting-started/providers/cloudflare#cloudflare-workers-ai-provider-guide","content":"Open-model inference at the edge via Cloudflare Workers AI","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Workers AI Provider Guide","lvl2":"Cloudflare Workers AI Provider Guide","lvl3":""}},
 {"objectID":"25503e78c2d92fb7cee70f0a5e3045de4dbad7279ef1b5cb1a6c834a89c4d006","title":"Overview","url":"/docs/getting-started/providers/cloudflare#overview","content":"Cloudflare Workers AI\nserves Meta Llama, Mistral, and other open models from Cloudflare's\nglobal GPU cluster. NeuroLink talks to the OpenAI-compatible endpoint.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Workers AI Provider Guide","lvl2":"Overview","lvl3":""}},
 {"objectID":"11f1e13f3ceb9f43bdaceb8669617ec44d6a62676514a7899f65463375681501","title":"Key Facts","url":"/docs/getting-started/providers/cloudflare#key-facts","content":"Protocol: OpenAI-compatible (/v1/chat/completions)\nDefault base URL: https://api.cloudflare.com/client/v4/accounts/{accountId}/ai/v1\nDefault model: @cf/meta/llama-3.3-70b-instruct-fp8-fast\nStreaming: Yes\nTool calling: Limited (model-dependent)","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Workers AI Provider Guide","lvl2":"Key Facts","lvl3":""}},
@@ -5683,7 +5687,7 @@
 {"objectID":"088e2cc83205e7cef6bcdc0633aee37efb875a215914292fe0d8472ad9956f7f","title":"Supported Models (sample)","url":"/docs/getting-started/providers/cloudflare#supported-models-sample","content":"| Model ID                                   | Notes          |\n| ------------------------------------------ | -------------- |\n| @cf/meta/llama-3.3-70b-instruct-fp8-fast | Default        |\n| @cf/meta/llama-3.1-70b-instruct          | Llama 3.1 70B  |\n| @cf/meta/llama-3.1-8b-instruct           | Fast tier      |\n| @cf/meta/llama-3.2-11b-vision-instruct   | Vision-capable |\n\nBrowse: https://developers.cloudflare.com/workers-ai/models","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Workers AI Provider Guide","lvl2":"Supported Models (sample)","lvl3":""}},
 {"objectID":"e1e8b466c22f042a3d3095b12f19a9fe24a96d6480b008310cfc189f341d45a5","title":"Provider Aliases","url":"/docs/getting-started/providers/cloudflare#provider-aliases","content":"| Alias        | Example                 |\n| ------------ | ----------------------- |\n| cloudflare | --provider cloudflare |\n| cf         | --provider cf         |","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Workers AI Provider Guide","lvl2":"Provider Aliases","lvl3":""}},
 {"objectID":"62de16314951fdc73f9dd06c4c146e4c9202ddb9445bea480242fcac6cf0ab83","title":"Configuration Reference","url":"/docs/getting-started/providers/cloudflare#configuration-reference","content":"| Environment Variable    | Required | Default                                    |\n| ----------------------- | -------- | ------------------------------------------ |\n| CLOUDFLARE_ACCOUNT_ID | Yes      | —                                          |\n| CLOUDFLARE_API_KEY    | Yes      | —                                          |\n| CLOUDFLARE_MODEL      | No       | @cf/meta/llama-3.3-70b-instruct-fp8-fast |","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Workers AI Provider Guide","lvl2":"Configuration Reference","lvl3":""}},
-{"objectID":"a12a3191a8590f0031973af941fe6c2be7f7cf931536ea38903ef8e8bf9bb5c6","title":"Also configures decisions","url":"/docs/getting-started/providers/cloudflare#also-configures-decisions","content":"CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID also configure a separate\nprovider, the Cloudflare Clef provider (cloudflare-clef),\nwhich serves decide() and emits no text. It is the second-to-last decision\nprovider NeuroLink falls back to: built-in features that call decide() use the\nfirst one configured, in the order TypeSafe, Laya, XOR, SubQ, Perplexity, Clef, then\nOpenAI Decisions. So a host\nthat set these two variables only for this text provider, and has no other\ndecision provider configured, has Clef as its default decision provider, and no\nswitch turns that off while they are set.\n\ncredentials.cloudflare does not configure decide; Clef reads its own\nslice, credentials.cloudflareClef. The text provider itself is unchanged: it\nstill serves generate() and stream() with the same variables, models and base\nURL. See\nOne token, two providers.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Workers AI Provider Guide","lvl2":"Also configures decisions","lvl3":""}},
+{"objectID":"a12a3191a8590f0031973af941fe6c2be7f7cf931536ea38903ef8e8bf9bb5c6","title":"Also configures decisions","url":"/docs/getting-started/providers/cloudflare#also-configures-decisions","content":"CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID also configure a separate\nprovider, the Cloudflare Clef provider (cloudflare-clef),\nwhich serves decide() and emits no text. It is the sixth of the eight decision\nproviders NeuroLink falls back to: built-in features that call decide() use the\nfirst one configured, in the order TypeSafe, Laya, XOR, SubQ, Perplexity, Clef,\nOpenAI Decisions, then ScaleDown. So a host\nthat set these two variables only for this text provider, and has no other\ndecision provider configured, has Clef as its default decision provider, and no\nswitch turns that off while they are set.\n\ncredentials.cloudflare does not configure decide; Clef reads its own\nslice, credentials.cloudflareClef. The text provider itself is unchanged: it\nstill serves generate() and stream() with the same variables, models and base\nURL. See\nOne token, two providers.","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Workers AI Provider Guide","lvl2":"Also configures decisions","lvl3":""}},
 {"objectID":"c18937da302797e2d1727a8d5894df09424bcf4d107a5500e2a04ce958eb9673","title":"See Also","url":"/docs/getting-started/providers/cloudflare#see-also","content":"Cloudflare Clef Provider — typed decide() judgments on the same token and account id\nTogether AI Provider\nFireworks Provider","hierarchy":{"lvl0":"Getting Started","lvl1":"Cloudflare Workers AI Provider Guide","lvl2":"See Also","lvl3":""}},
 {"objectID":"771db3571626e1c1eb3472bc9a61338f884160a5415735e50731de5e772d00f0","title":"Cohere Provider Guide","url":"/docs/getting-started/providers/cohere","content":"Cohere Provider Guide\n\nCommand R chat + Embed v3 embeddings via the Cohere API\n\nOverview\n\nCohere offers a production-grade chat\ncatalog (Command R / R+ / R7B) plus top-tier embeddings (Embed v3) and\nreranking (Rerank v3). NeuroLink wraps chat via the OpenAI-compatible\nendpoint and embeddings via the native /v2/embed endpoint.\n\nKey Facts\nProtocol: OpenAI-compatible chat at /compatibility/v1,\n  native embed at /v2/embed\nDefault base URL: https://api.cohere.com/compatibility/v1\nDefault chat model: command-r-plus-08-2024\nDefault embed model: embed-english-v3.0\nStreaming: Yes\nTool calling: Yes (Command R / R+)\n\nQuick Start\nGet an API Key\n\nhttps://dashboard.cohere.com/api-keys\nConfigure\nGenerate Text\nGenerate Embeddings\n\nSupported Models\n\n| Model ID                      | Family           | Notes                    |\n| ----------------------------- | ---------------- | ------------------------ |\n| command-r-plus-08-2024      | Chat (default)   | Flagship                 |\n| command-r-08-2024           | Chat             | Mid-tier                 |\n| command-r7b-12-2024         | Chat             | Most compact             |\n| command-a-reasoning-08-2025 | Reasoning        | Reasoning traces         |\n| embed-english-v3.0          | Embeddings (def) | 1024 dim, English        |\n| embed-multilingual-v3.0     | Embeddings       | 1024 dim, 100+ languages |\n\nCLI Usage\n\nConfiguration Reference\n\n| Environment Variable | Required | Default                                   |\n| -------------------- | -------- | ----------------------------------------- |\n| COHERE_API_KEY     | Yes      | —                                         |\n| COHERE_MODEL       | No       | command-r-plus-08-2024                  |\n| COHERE_BASE_URL    | No       | https://api.cohere.com/compatibility/v1 |\n\nFeature Support Matrix\n\n| Feature           | Support      |\n| ----------------- | ------------ |\n| Text generation   | Yes          |\n| Streaming         | Yes          |\n| Tool calling      | Yes          |\n| Structured output | Yes          |\n| Vision            | No           |\n| Embeddings        | Yes (native) |\n| Reranking         | Yes          |\n\nTroubleshooting\nNot Found — the chosen model may not be available on your tier.\n  Try command-r-plus-08-2024 or command-r-08-2024.\n\nSee Also\nVoyage Provider\nJina Provider","hierarchy":{"lvl0":"Getting Started","lvl1":"Cohere Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"765ab44d68aaf7fc0fe06a519c53fb7aad3106753d2a1b679c26293cce92e958","title":"Cohere Provider Guide","url":"/docs/getting-started/providers/cohere#cohere-provider-guide","content":"Command R chat + Embed v3 embeddings via the Cohere API","hierarchy":{"lvl0":"Getting Started","lvl1":"Cohere Provider Guide","lvl2":"Cohere Provider Guide","lvl3":""}},
@@ -6232,14 +6236,15 @@
 {"objectID":"6ad36ba3488491bb913c7a77b86a278cc86509da88a054975bcded71f5c02130","title":"[OpenRouter](/docs/getting-started/providers/openrouter)","url":"/docs/getting-started/providers#openrouterdocsgetting-startedprovidersopenrouter","content":"300+ models from 60+ providers\n🌐 Single API across many AI providers (Anthropic, OpenAI, Google, Meta, etc.)\n⚡ Automatic failover and routing\n💰 Competitive pricing with cost optimization\n🎯 Zero lock-in - switch models instantly\n📊 Usage tracking dashboard\n🆓 Free models available\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[OpenRouter](/docs/getting-started/providers/openrouter)","lvl3":""}},
 {"objectID":"733a2fa8ea4c547cdcdf502f89f003a2208c0e5828e36e7faaaa94e0903d3d49","title":"[OpenAI Compatible](/docs/getting-started/providers/openai-compatible)","url":"/docs/getting-started/providers#openai-compatibledocsgetting-startedprovidersopenai-compatible","content":"OpenRouter, vLLM, LocalAI, and more\n🌐 100+ models through OpenRouter\n💻 Local deployment with vLLM\n🔓 Self-hosted with LocalAI\n🔄 Drop-in OpenAI replacement\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[OpenAI Compatible](/docs/getting-started/providers/openai-compatible)","lvl3":""}},
 {"objectID":"fc415454bfde58f3634df544ef2adb88836c75fb1344b7042af7a9d6fdfa63d4","title":"[LiteLLM](/docs/getting-started/providers/litellm)","url":"/docs/getting-started/providers#litellmdocsgetting-startedproviderslitellm","content":"100+ providers through proxy\n🔄 Unified API for 100+ providers\n📊 Load balancing and fallbacks\n💰 Cost tracking\n🎯 Model routing\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[LiteLLM](/docs/getting-started/providers/litellm)","lvl3":""}},
-{"objectID":"6337b6964b5ebadd9610f93633594d912ee39f49c2ef5e1683214dce45a1aee3","title":"🧠 Decision-Only Providers","url":"/docs/getting-started/providers#decision-only-providers","content":"The providers that serve decide rather than generate/stream:\nTypeSafe, Laya, XOR,\nSubQ, Perplexity, Cloudflare Clef and\nOpenAI Decisions. Each returns typed boolean/choice/score\nanswers and emits no text, so none appears in generation fallback chains or the\nhealth sweep.","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"🧠 Decision-Only Providers","lvl3":""}},
+{"objectID":"6337b6964b5ebadd9610f93633594d912ee39f49c2ef5e1683214dce45a1aee3","title":"🧠 Decision-Only Providers","url":"/docs/getting-started/providers#decision-only-providers","content":"The providers that serve decide rather than generate/stream:\nTypeSafe, Laya, XOR,\nSubQ, Perplexity, Cloudflare Clef,\nOpenAI Decisions and ScaleDown. Each returns typed boolean/choice/score\nanswers and emits no text, so none appears in generation fallback chains or the\nhealth sweep.","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"🧠 Decision-Only Providers","lvl3":""}},
 {"objectID":"50860d1f7a27dbe3ad7452ad7079ea437658db372b97081d0a509429369cacc6","title":"[TypeSafe (Jev)](/docs/getting-started/providers/typesafe)","url":"/docs/getting-started/providers#typesafe-jevdocsgetting-startedproviderstypesafe","content":"Typed, calibrated judgments instead of text\n🎯 boolean / choice / score answers, each with a calibrated confidence\n⚡ Latency flat in question count — 1 question ~393 ms, 400 questions ~465 ms\n💰 ~$0.00002 per decision (~$0.042/M input, output billed at zero)\n🔌 Two transports: TypeSafe direct, or the Vercel AI Gateway\n🛡️ Fails open — with no decision provider configured, every consumer behaves exactly as before\n🔑 API key from console.typesafe.ai/keys\n🔄 Aliases: jev, typesafe-ai\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[TypeSafe (Jev)](/docs/getting-started/providers/typesafe)","lvl3":""}},
 {"objectID":"43b838d60f2794ce46361196d6fc8a708b3a301398dd55f2c8015d567aa21c2c","title":"[Laya](/docs/getting-started/providers/laya)","url":"/docs/getting-started/providers#layadocsgetting-startedproviderslaya","content":"Open-weights decision provider — the same typed boolean/choice/score answers as Jev, from Convai Innovations' Apache-2.0 checkpoints, at a Laya server or LiteLLM proxy route you configure\n🧭 Serves decide only; built-in features use it when its key and base URL are set and neither TYPESAFE_API_KEY nor AI_GATEWAY_API_KEY is\n📏 Refuses more than ~768 tokens of state (320 on english) before any network call, since its encoders read only 1,024 (512)\n🔌 No built-in endpoint: LAYA_BASE_URL (or credentials.laya.baseURL) names a Laya server or a LiteLLM pass-through route to one\n🔑 LAYA_API_KEY is the key that endpoint accepts; on LiteLLM, the route must be in the key's Allowed Routes\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[Laya](/docs/getting-started/providers/laya)","lvl3":""}},
 {"objectID":"613cbeca5e8fa0bbe765b8db9ec4ed2ff76a913dd45b8dab5e2de06a6e5f19ff","title":"[XOR](/docs/getting-started/providers/xor)","url":"/docs/getting-started/providers#xordocsgetting-startedprovidersxor","content":"Open-weights decision provider — the same typed boolean/choice/score answers as Jev, from xor-1.1, Juspay's Apache-2.0 model, at a deployment or LiteLLM proxy route you configure\n🧭 Serves decide only; built-in features use it when its key and base URL are set and no TypeSafe key, no AI_GATEWAY_API_KEY and no Laya key and base URL is\n🖼️ Takes up to 8 images or one video with a decision, as a Buffer, a local file path or a data: URL; TypeSafe and Laya refuse media before any request, and Perplexity refuses a video\n📏 Refuses more than about 200,000 estimated tokens of state before any network call\n🔌 No built-in endpoint: XOR_BASE_URL (or credentials.xor.baseURL) is the origin of a deployment or of a LiteLLM route to one; requests go to <base URL>/v1/systemone\n🔑 XOR_API_KEY is the key that endpoint accepts; on LiteLLM, the key's team must allow xor-1.1\n📦 Weights and setup: huggingface.co/juspay/xor\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[XOR](/docs/getting-started/providers/xor)","lvl3":""}},
 {"objectID":"9fd996aaa27fb32250d5434f0e2764d8bf348370af0e011f9e447a0fdf26e3a8","title":"[SubQ](/docs/getting-started/providers/subq)","url":"/docs/getting-started/providers#subqdocsgetting-startedproviderssubq","content":"Hosted decision provider — the same typed boolean/choice/score answers as Jev, from Subquadratic's subq-snap-preview (public preview), at https://api.subq.ai\n🧭 Serves decide only (provider id subq; the aliases subquadratic and subq-snap work too); built-in features use it when it is configured and none of TypeSafe, Laya or XOR is — it comes ahead of Perplexity and Cloudflare Clef, because its key is shared with no other provider\n🔑 SUBQ_API_KEY alone configures it (keys come from console.subq.ai, shown once); SUBQ_MODEL is optional\n🖼️ No image or video input yet: SubQ's image format is not one the decide catalog can describe, so images are refused before any request\n📏 NeuroLink reports an advisory state window (376,000 estimated tokens; SubQ's own 400 response states a maximum model length of 393,216) and enforces none; no question cap is declared\n🔌 Hosted endpoint, so no base URL is needed; SUBQ_BASE_URL (or credentials.subq.baseURL) can name another origin\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[SubQ](/docs/getting-started/providers/subq)","lvl3":""}},
 {"objectID":"6ab8cdd8152a0d7c9844d54b6d4eaccc905b60e79f4b28508d78d52c3cfe9470","title":"[Perplexity Decisions](/docs/getting-started/providers/perplexity-decider)","url":"/docs/getting-started/providers#perplexity-decisionsdocsgetting-startedprovidersperplexity-decider","content":"Hosted decision provider — the same typed boolean/choice/score answers as Jev, from Perplexity's pplx-decider-v1-27b, at https://api.perplexity.ai\n🧭 Serves decide only (provider id perplexity-decider, not the Sonar text provider perplexity); built-in features use it when it is configured and no TypeSafe key, no AI_GATEWAY_API_KEY, no Laya key and base URL, no XOR key and base URL and no SUBQ_API_KEY is\n🔑 PERPLEXITY_API_KEY alone configures it, and it is the same key the Perplexity text provider reads, so a key set for Sonar also lets built-in features use it when none of TypeSafe, Laya, XOR or SubQ is configured; the guide lists what each consumer then sends\n🖼️ Takes up to 8 PNG, JPEG or WebP images with a decision; no video\n💰 $0.04 per million input tokens (image tokens included), output free, as Perplexity documents it\n📏 Refuses more than about 100,000 estimated tokens of state (NeuroLink's own window; the server's ceiling is 262,144 input tokens), or more than 128 questions in a decide() call (tryDecide() splits them), before any network call. The 128-question cap, the server's token ceiling and the latency were measured on a real account in October 2026\n🔌 Hosted endpoint, so no base URL is needed; PERPLEXITY_DECIDER_BASE_URL (or credentials.perplexityDecider.baseURL) can name another origin\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[Perplexity Decisions](/docs/getting-started/providers/perplexity-decider)","lvl3":""}},
-{"objectID":"94b24c75db64d3f04ba04b2ee0f04ec13bc0509de0aea1fa1b3e5516dea73ea9","title":"[Cloudflare Clef](/docs/getting-started/providers/cloudflare-clef)","url":"/docs/getting-started/providers#cloudflare-clefdocsgetting-startedproviderscloudflare-clef","content":"Hosted decision provider on Workers AI — the same typed boolean/choice/score answers as Jev, from Cloudflare's clef (27B) and clef-flash (9B), at https://api.cloudflare.com/client/v4\n🧭 Serves decide only (provider id cloudflare-clef, not the Workers AI text provider cloudflare); built-in features use it only when it is configured and none of TypeSafe, Laya, XOR, SubQ or Perplexity is — it comes after those five\n🔑 CLOUDFLARE_API_KEY (Workers AI permission) and CLOUDFLARE_ACCOUNT_ID configure it, and they are the same two variables the Cloudflare text provider reads, so a pair set for that provider also lets built-in features use it when no other decision provider is configured\n🖼️ Takes up to 4 PNG, JPEG or WebP images with a decision; no video\n💰 $0.24 per million input tokens for clef, $0.09 for clef-flash; no output price is listed on Cloudflare's pricing page\n📏 The state window changed in October 2026 (hosted service or model: unknown): until 2026-10-04 the endpoint ignored text past about 2,048 tokens without an error, and on 2026-10-07 it read states of at least 190,153 tokens (clef-flash) and 52,154 (clef); NeuroLink refuses more than 16,000 estimated tokens (digits count a token each, punctuation 0.75, emoji 3), more than 64 questions in a decide() call (tryDecide() splits them) and a request over 256,000 bytes, base64 image data included, before any network call. The state cut and the request ceiling were measured on a real account in October 2026, with the follow-up text cuts and API limits checked on both models; see the guide for the tested shapes and dates\n🔌 Hosted endpoint, so no base URL is needed; CLOUDFLARE_CLEF_BASE_URL (or credentials.cloudflareClef.baseURL) can name another base\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[Cloudflare Clef](/docs/getting-started/providers/cloudflare-clef)","lvl3":""}},
-{"objectID":"bcc625210e66063fe8068d952207922cf92704dbb5424a7a3f8192f66de3bab2","title":"[OpenAI Decisions](/docs/getting-started/providers/openai-decisions)","url":"/docs/getting-started/providers#openai-decisionsdocsgetting-startedprovidersopenai-decisions","content":"Hosted decision provider (public beta) — the same typed boolean/choice/score answers as Jev, from OpenAI's Decisions API (gpt-6-luna), at https://api.openai.com/v1/decisions\n🧭 Serves decide only (provider id openai-decisions, not the OpenAI chat provider openai); built-in features use it only when it is configured and none of TypeSafe, Laya, XOR, SubQ, Perplexity or Cloudflare Clef is — it comes last\n🔑 OPENAI_API_KEY alone configures it, and it is the same key the OpenAI chat provider reads, so a key set for chat also lets built-in features (and a transcription dictionary's correction check) use it, and make paid decision calls, when no other decision provider is configured, unless OPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com (the key is then taken to be a gateway's and is not used by default; a call that names the provider still uses it and sends it to api.openai.com); the guide lists what each consumer then sends\n🖼️ Takes up to 8 images as inline data URLs with a decision (NeuroLink's own allowance, not OpenAI's); no video\n💰 $0.10 per million input tokens, input only, per OpenAI's Decisions guide as read on 2026-10-07; regional and long-context multipliers are not modelled\n📏 Question and option limits, image formats, rate limits and latency are not established (none measured); NeuroLink reports 922,000 state tokens as an advisory figure and never refuses on it\n🔌 Hosted endpoint, so no base URL is needed; OPENAI_DECISIONS_BASE_URL (or credentials.openaiDecisions.baseURL) can name another origin\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[OpenAI Decisions](/docs/getting-started/providers/openai-decisions)","lvl3":""}},
+{"objectID":"94b24c75db64d3f04ba04b2ee0f04ec13bc0509de0aea1fa1b3e5516dea73ea9","title":"[Cloudflare Clef](/docs/getting-started/providers/cloudflare-clef)","url":"/docs/getting-started/providers#cloudflare-clefdocsgetting-startedproviderscloudflare-clef","content":"Hosted decision provider on Workers AI — the same typed boolean/choice/score answers as Jev, from Cloudflare's clef (27B) and clef-flash (9B), at https://api.cloudflare.com/client/v4\n🧭 Serves decide only (provider id cloudflare-clef, not the Workers AI text provider cloudflare); built-in features use it only when it is configured and none of TypeSafe, Laya, XOR, SubQ or Perplexity is — it comes after those five\n🔑 CLOUDFLARE_API_KEY (Workers AI permission) and CLOUDFLARE_ACCOUNT_ID configure it, and they are the same two variables the Cloudflare text provider reads, so a pair set for that provider also lets built-in features use it when none of TypeSafe, Laya, XOR, SubQ or Perplexity is configured\n🖼️ Takes up to 4 PNG, JPEG or WebP images with a decision; no video\n💰 $0.24 per million input tokens for clef, $0.09 for clef-flash; no output price is listed on Cloudflare's pricing page\n📏 The state window changed in October 2026 (hosted service or model: unknown): until 2026-10-04 the endpoint ignored text past about 2,048 tokens without an error, and on 2026-10-07 it read states of at least 190,153 tokens (clef-flash) and 52,154 (clef); NeuroLink refuses more than 16,000 estimated tokens (digits count a token each, punctuation 0.75, emoji 3), more than 64 questions in a decide() call (tryDecide() splits them) and a request over 256,000 bytes, base64 image data included, before any network call. The state cut and the request ceiling were measured on a real account in October 2026, with the follow-up text cuts and API limits checked on both models; see the guide for the tested shapes and dates\n🔌 Hosted endpoint, so no base URL is needed; CLOUDFLARE_CLEF_BASE_URL (or credentials.cloudflareClef.baseURL) can name another base\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[Cloudflare Clef](/docs/getting-started/providers/cloudflare-clef)","lvl3":""}},
+{"objectID":"bcc625210e66063fe8068d952207922cf92704dbb5424a7a3f8192f66de3bab2","title":"[OpenAI Decisions](/docs/getting-started/providers/openai-decisions)","url":"/docs/getting-started/providers#openai-decisionsdocsgetting-startedprovidersopenai-decisions","content":"Hosted decision provider (public beta) — the same typed boolean/choice/score answers as Jev, from OpenAI's Decisions API (gpt-6-luna), at https://api.openai.com/v1/decisions\n🧭 Serves decide only (provider id openai-decisions, not the OpenAI chat provider openai); built-in features use it only when it is configured and none of TypeSafe, Laya, XOR, SubQ, Perplexity or Cloudflare Clef is — it comes seventh, ahead of only ScaleDown\n🔑 OPENAI_API_KEY alone configures it, and it is the same key the OpenAI chat provider reads, so a key set for chat also lets built-in features (and a transcription dictionary's correction check) use it, and make paid decision calls, when none of the six providers ahead of it is configured, unless OPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com (the key is then taken to be a gateway's and is not used by default; a call that names the provider still uses it and sends it to api.openai.com); the guide lists what each consumer then sends\n🖼️ Takes up to 8 images as inline data URLs with a decision (NeuroLink's own allowance, not OpenAI's); no video\n💰 $0.10 per million input tokens, input only, per OpenAI's Decisions guide as read on 2026-10-07; regional and long-context multipliers are not modelled\n📏 Question and option limits, image formats, rate limits and latency are not established (none measured); NeuroLink reports 922,000 state tokens as an advisory figure and never refuses on it\n🔌 Hosted endpoint, so no base URL is needed; OPENAI_DECISIONS_BASE_URL (or credentials.openaiDecisions.baseURL) can name another origin\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[OpenAI Decisions](/docs/getting-started/providers/openai-decisions)","lvl3":""}},
+{"objectID":"900a7065e09e85ae5c74d7f57d76f6ca01ac8e2ab21e947eab38b01314b0cb52","title":"[ScaleDown](/docs/getting-started/providers/scaledown)","url":"/docs/getting-started/providers#scaledowndocsgetting-startedprovidersscaledown","content":"Hosted decision provider — the same typed boolean/choice/score answers as Jev, from ScaleDown's Classify models (classify-1 by default), at https://api.scaledown.xyz/v1/scaledown\n🧭 Serves decide only (provider id scaledown; the alias scaledown-classify works too); built-in features use it only when it is configured and none of TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef or OpenAI Decisions is — it comes last, by choice and not because its key is shared\n🔑 SCALEDOWN_API_KEY alone configures it, and no other provider reads that variable; the key is sent in an x-api-key header, never as a Bearer token\n🖼️ Text only: images and video are refused before any request, and the state is always sent as a string (a string unchanged, anything else as its JSON text; a blank string is refused locally)\n💰 USD 0.05 per million input tokens, output not charged: a list price reported by the marketing session on 2026-10-15, not measured here. A choice question is scored once per option, so input tokens grow with options times state length\n📏 Context size, latency, rate limits and the option, question and level maximums are not measured; no state window is declared and nothing is refused on size\n🔌 Hosted endpoint, so no base URL is needed; SCALEDOWN_BASE_URL (or credentials.scaledown.baseURL) can name an origin or an origin with a path prefix\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[ScaleDown](/docs/getting-started/providers/scaledown)","lvl3":""}},
 {"objectID":"c98403894a115164f29a019817b7f39c93b59623a9854b4575ad06bc5f597320","title":"🧩 Additional Catalog Providers","url":"/docs/getting-started/providers#-additional-catalog-providers","content":"Every provider below is a Tier-2 catalog entry. Provider-specific\nbehavioural quirks may apply, so check the provider's catalog file. The whole\nintegration is one JSON file under src/lib/providers/catalog/. Each page is\ngenerated from that file, which is also what the CI onboarding gate reads.","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"🧩 Additional Catalog Providers","lvl3":""}},
 {"objectID":"026138da0a559269ecd9f0b2ceb72d9039096e72e064142afb7bfb74a5cb7455","title":"[API Route](/docs/getting-started/providers/api-route)","url":"/docs/getting-started/providers#api-routedocsgetting-startedprovidersapi-route","content":"Claude Sonnet 4.6\n🤖 8 models; default claude-sonnet-4-6 (1M context)\n🛠️ Native tool calling + structured output together\n💳 Free tier available\n✅ Roster verified 2026-09-17 (authenticated GET /v1/models)\n🔑 API key from api-route.com\n🔄 Aliases: apiroute\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[API Route](/docs/getting-started/providers/api-route)","lvl3":""}},
 {"objectID":"85aa65949d959db077347ecf93a4c40ade9ebab4300d4d52c62c702f6929053b","title":"[Baseten](/docs/getting-started/providers/baseten)","url":"/docs/getting-started/providers#basetendocsgetting-startedprovidersbaseten","content":"GLM 5.3 Flash\n🤖 16 models; default zai-org/GLM-5.3-Flash (1M context)\n🛠️ Native tool calling + structured output together\n💳 Free tier available\n✅ Roster verified 2026-09-03 (authenticated GET /v1/models)\n🔑 API key from app.baseten.co\n\nSetup Guide →","hierarchy":{"lvl0":"Getting Started","lvl1":"AI Provider Guides","lvl2":"[Baseten](/docs/getting-started/providers/baseten)","lvl3":""}},
@@ -6326,18 +6331,18 @@
 {"objectID":"4e0264790a48f8469b5d25d609cab04fae1ad3ba2ccee40c8829cd8571c265c6","title":"Verification status","url":"/docs/getting-started/providers/koscompute#verification-status","content":"Tier-2 onboarding requires evidence before a provider is accepted, and\npnpm run verify:provider-onboarding gates it in CI. This is what the\ncatalog records for KosCompute — and, just as importantly, what it does\nnot yet record:\n\n| Probe                 | Result                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |\n| --------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Roster                | unauthenticated GET /v1/models, HTTP 200, 7 models, 2026-09-28. No API key was used or required for this call.                                                                                                                                                                                                                                                                                                                                                                                    |\n| Auth rejection        | Not probed. Would require a POST with an invalid key, which this credential-free onboarding pass does not send.","hierarchy":{"lvl0":"Getting Started","lvl1":"KosCompute Provider Guide","lvl2":"Verification status","lvl3":""}},
 {"objectID":"3944867506991deef1761c0dab4d9fa3bc0efe0d53bfb3f9acb78437c41f3ecf","title":"Troubleshooting","url":"/docs/getting-started/providers/koscompute#troubleshooting","content":"| Symptom                              | Cause                                                                   | Fix                                                                                                                                                  |\n| ------------------------------------ | ----------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Invalid KosCompute API key         | KOSCOMPUTE_API_KEY unset, wrong, expired, or lacks access             | KosCompute's docs describe 401 as \"missing, invalid, expired, inactive, or conflicting credentials\" but publish no example error body or code string |\n| Model not found                      | Wrong model id, or the roster changed since 2026-09-28                  | Pick a current id from an unauthenticated GET https://api.koscompute.com/v1/models                                                                 |\n| 429 with concurrent_limit_exceeded | Per-key concurrency limit exceeded                                      | Wait for an in-flight request to finish, or reduce parallel requests, before retrying                                                                |\n| No signup page found                 | Expected — KosCompute publishes no public signup or key-management page | Obtain a key out of band; there is no self-serve console to link to                                                                                  |","hierarchy":{"lvl0":"Getting Started","lvl1":"KosCompute Provider Guide","lvl2":"Troubleshooting","lvl3":""}},
 {"objectID":"2de12b046642ad2cef71f2aa84ab6de9163a41b51a046794c2d34738958d382d","title":"See also","url":"/docs/getting-started/providers/koscompute#see-also","content":"Provider setup overview\nAll providers\nTier-2 onboarding — how this provider's JSON becomes a working integration\nProvider feature compatibility","hierarchy":{"lvl0":"Getting Started","lvl1":"KosCompute Provider Guide","lvl2":"See also","lvl3":""}},
-{"objectID":"6d919d5032d8bf60b8341d6b8ff90eeb0c7d2f87553ca6e6e96d08b988c7e6e9","title":"Laya Provider Guide","url":"/docs/getting-started/providers/laya","content":"Laya Provider Guide\n\nA provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, from an open-weights model you can\nrun yourself. It emits no text at all.\n\nOverview\n\nLaya is Convai Innovations' Apache-2.0 \"System One\" decision model. You send\none state plus named, typed questions; an encoder answers every question in a\nsingle forward pass and returns a probability distribution for each. NeuroLink\ncalls whatever base URL you configure — a Laya server, or a LiteLLM proxy with a\npass-through route to one. There is no built-in endpoint.\n\nKey Facts\n\n|                       |                                                                                                   |\n| --------------------- | ------------------------------------------------------------------------------------------------- |\n| Inference type        | decide only                                                                                     |\n| Checkpoints           | typed-decisions (default, fine-tuned for typed decisions), multilingual, english, or auto |\n| Input window          | 1,024 tokens (typed-decisions, multilingual); 512 (english)                                 |\n| Questions per request | up to 64                                                                                          |\n| Endpoint              | <base URL>/predict; the base URL is required, with no default                                   |\n| Precedence            | used automatically when its key and base URL are set and no TypeSafe key is                       |\n\nQuick Start\nGet an endpoint and a key\n\nRun a Laya server (see below), or use a LiteLLM\nproxy with a pass-through route to one. On LiteLLM, create a virtual key and add\nthe route's /predict path (and /health, if you want to probe it) to the\nkey's Allowed Routes; a key without them is refused with a 403\nKey/team not allowed to access passthrough route.\nConfigure\n\nBoth the base URL and the key are required. Set them in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:\nUse it\n\nFrom the CLI\n\nWhen NeuroLink uses it\n\nEvery built-in consumer of decide — model routing, relevance-driven\ncompaction, tool routing and RAG planning — asks for the default decision\nprovider. That is the first one that is configured — in the environment or in\nthe credentials passed to the SDK — in the order TypeSafe, Laya,\nXOR, SubQ, Perplexity,\nCloudflare Clef, then\nOpenAI Decisions. TypeSafe has two keys,\nTYPESAFE_API_KEY and AI_GATEWAY_API_KEY (its Vercel AI Gateway route), and\neither one counts. Laya counts only with both its key and its base URL, and so\ndoes XOR. SubQ counts with its key alone, SUBQ_API_KEY. Perplexity counts with\nits key alone, and that key,\nPERPLEXITY_API_KEY, is shared with Perplexity's text provider. Cloudflare Clef\ncounts only with both CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID, the same\ntwo variables the Workers AI text provider reads. OpenAI Decisions counts with its\nkey alone, OPENAI_API_KEY, which the OpenAI chat provider reads too. So:\nA TypeSafe key, plus Laya's key and base URL: built-in features use\n  TypeSafe; Laya runs only where a caller asks for provider: \"laya\".\nOnly Laya's key and base URL: built-in features use Laya.\nLaya's key and base URL, plus XOR's: built-in features use Laya; XOR runs\n  only where a caller asks for provider: \"xor\".\nA Laya key with no base URL: Laya is not configured. Built-in features\n  ignore it, and provider: \"laya\" fails with Laya requires a base URL.\nLaya's key and base URL, plus a Perplexity key: built-in features use\n  Laya; Perplexity runs only where a caller asks for\n  provider: \"perplexity-decider\".\nLaya's key and base URL, plus CLOUDFLARE_API_KEY and\n  CLOUDFLARE_ACCOUNT_ID: built-in features use Laya; Cloudflare Clef runs\n  only where a caller asks for provider: \"cloudflare-clef\".\nLaya's key and base URL, plus SUBQ_API_KEY: built-in features use Laya;\n  SubQ runs only where a caller asks for provider: \"subq\".\nNone of TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef or OpenAI\n  Decisions: everything behaves exactly as it did without a decision model.\n\nLimits\n\nThe window is small. Laya's encoders read 1,024 tokens, and 512 on the\nenglish checkpoint, against roughly 33,000 for Jev. Part of that is reserved\nfor each question and its options, so NeuroLink allows about 768 tokens of\nstate on typed-decisions and multilingual, and 320 on english, auto and\nany model name it does not recognise. Laya's server does not refuse a longer\nstate — it answers from the start of it and says nothing — so NeuroLink\nrefuses it instead, before any network call, with max_tokens_exceeded.\n\nThe size is an estimate, not Laya's tokenizer: about four characters per token\nfor ASCII text, and 1.5 tokens per character for other scripts (0.6 on\nmultilingual), calibrated against a live Laya 0.3.5 server. It errs toward refusing. Bu","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"","lvl3":""}},
+{"objectID":"6d919d5032d8bf60b8341d6b8ff90eeb0c7d2f87553ca6e6e96d08b988c7e6e9","title":"Laya Provider Guide","url":"/docs/getting-started/providers/laya","content":"Laya Provider Guide\n\nA provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, from an open-weights model you can\nrun yourself. It emits no text at all.\n\nOverview\n\nLaya is Convai Innovations' Apache-2.0 \"System One\" decision model. You send\none state plus named, typed questions; an encoder answers every question in a\nsingle forward pass and returns a probability distribution for each. NeuroLink\ncalls whatever base URL you configure — a Laya server, or a LiteLLM proxy with a\npass-through route to one. There is no built-in endpoint.\n\nKey Facts\n\n|                       |                                                                                                   |\n| --------------------- | ------------------------------------------------------------------------------------------------- |\n| Inference type        | decide only                                                                                     |\n| Checkpoints           | typed-decisions (default, fine-tuned for typed decisions), multilingual, english, or auto |\n| Input window          | 1,024 tokens (typed-decisions, multilingual); 512 (english)                                 |\n| Questions per request | up to 64                                                                                          |\n| Endpoint              | <base URL>/predict; the base URL is required, with no default                                   |\n| Precedence            | used automatically when its key and base URL are set and no TypeSafe key is                       |\n\nQuick Start\nGet an endpoint and a key\n\nRun a Laya server (see below), or use a LiteLLM\nproxy with a pass-through route to one. On LiteLLM, create a virtual key and add\nthe route's /predict path (and /health, if you want to probe it) to the\nkey's Allowed Routes; a key without them is refused with a 403\nKey/team not allowed to access passthrough route.\nConfigure\n\nBoth the base URL and the key are required. Set them in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:\nUse it\n\nFrom the CLI\n\nWhen NeuroLink uses it\n\nEvery built-in consumer of decide — model routing, relevance-driven\ncompaction, tool routing and RAG planning — asks for the default decision\nprovider. That is the first one that is configured — in the environment or in\nthe credentials passed to the SDK — in the order TypeSafe, Laya,\nXOR, SubQ, Perplexity,\nCloudflare Clef,\nOpenAI Decisions, then ScaleDown. TypeSafe has two keys,\nTYPESAFE_API_KEY and AI_GATEWAY_API_KEY (its Vercel AI Gateway route), and\neither one counts. Laya counts only with both its key and its base URL, and so\ndoes XOR. SubQ counts with its key alone, SUBQ_API_KEY. Perplexity counts with\nits key alone, and that key,\nPERPLEXITY_API_KEY, is shared with Perplexity's text provider. Cloudflare Clef\ncounts only with both CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID, the same\ntwo variables the Workers AI text provider reads. OpenAI Decisions counts with its\nkey alone, OPENAI_API_KEY, which the OpenAI chat provider reads too.\nScaleDown counts with its key alone, SCALEDOWN_API_KEY, which no other provider\nreads. So:\nA TypeSafe key, plus Laya's key and base URL: built-in features use\n  TypeSafe; Laya runs only where a caller asks for provider: \"laya\".\nOnly Laya's key and base URL: built-in features use Laya.\nLaya's key and base URL, plus XOR's: built-in features use Laya; XOR runs\n  only where a caller asks for provider: \"xor\".\nA Laya key with no base URL: Laya is not configured. Built-in features\n  ignore it, and provider: \"laya\" fails with Laya requires a base URL.\nLaya's key and base URL, plus a Perplexity key: built-in features use\n  Laya; Perplexity runs only where a caller asks for\n  provider: \"perplexity-decider\".\nLaya's key and base URL, plus CLOUDFLARE_API_KEY and\n  CLOUDFLARE_ACCOUNT_ID: built-in features use Laya; Cloudflare Clef runs\n  only where a caller asks for provider: \"cloudflare-clef\".\nLaya's key and base URL, plus SUBQ_API_KEY: built-in features use Laya;\n  SubQ runs only where a caller asks for provider: \"subq\".\nLaya's key and base URL, plus SCALEDOWN_API_KEY: built-in features use\n  Laya; ScaleDown runs only where a caller asks for provider: \"scaledown\".\nNone of TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef, OpenAI\n  Decisions or ScaleDown: everything behaves exactly as it did without a decision model.\n\nLimits\n\nThe window is small. Laya's encoders read 1,024 tokens, and 512 on the\nenglish checkpoint, against roughly 33,000 for Jev. Part of that is reserved\nfor each question and its options, so NeuroLink allows about 768 tokens of\nstate on typed-decisions and multilingual, and 320 on english, auto and\nany model name it does not recognise. Laya's server does not refuse a longer\nstate — it answers from the start of it and says nothing — so NeuroLink\nrefuses it instead, before any network call, with max_","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"c57a78b27b2b64f684ee60c36022d5b19a81323647a32687608b35530b4d3ed7","title":"Laya Provider Guide","url":"/docs/getting-started/providers/laya#laya-provider-guide","content":"A provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, from an open-weights model you can\nrun yourself. It emits no text at all.","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"Laya Provider Guide","lvl3":""}},
 {"objectID":"969ae83c9e8333da8d49838b0595db3588d8828b67611944cd46f08435f19d97","title":"Overview","url":"/docs/getting-started/providers/laya#overview","content":"Laya is Convai Innovations' Apache-2.0 \"System One\" decision model. You send\none state plus named, typed questions; an encoder answers every question in a\nsingle forward pass and returns a probability distribution for each. NeuroLink\ncalls whatever base URL you configure — a Laya server, or a LiteLLM proxy with a\npass-through route to one. There is no built-in endpoint.","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"Overview","lvl3":""}},
 {"objectID":"acf4d054e73fb41ef18db1c578cf1beb1117c3f43b24c6743ba30a4fe6569fb7","title":"Key Facts","url":"/docs/getting-started/providers/laya#key-facts","content":"|                       |                                                                                                   |\n| --------------------- | ------------------------------------------------------------------------------------------------- |\n| Inference type        | decide only                                                                                     |\n| Checkpoints           | typed-decisions (default, fine-tuned for typed decisions), multilingual, english, or auto |\n| Input window          | 1,024 tokens (typed-decisions, multilingual); 512 (english)                                 |\n| Questions per request | up to 64                                                                                          |\n| Endpoint              | <base URL>/predict; the base URL is required, with no default                                   |\n| Precedence            | used automatically when its key and base URL are set and no TypeSafe key is                       |","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"Key Facts","lvl3":""}},
 {"objectID":"ae5970b1821e7485e9d952384e4d6749a99d486ea648241bcbd8a9d43579e79a","title":"1. Get an endpoint and a key","url":"/docs/getting-started/providers/laya#1-get-an-endpoint-and-a-key","content":"Run a Laya server (see below), or use a LiteLLM\nproxy with a pass-through route to one. On LiteLLM, create a virtual key and add\nthe route's /predict path (and /health, if you want to probe it) to the\nkey's Allowed Routes; a key without them is refused with a 403\nKey/team not allowed to access passthrough route.","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"1. Get an endpoint and a key","lvl3":""}},
 {"objectID":"662bbcdcfe03b7cb5abd93bb0088ff4bdcd8558ad74a79d70f1f7ab36023842c","title":"2. Configure","url":"/docs/getting-started/providers/laya#2-configure","content":"Both the base URL and the key are required. Set them in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"2. Configure","lvl3":""}},
-{"objectID":"c99b86a6719e50f022a261b2688e6c849e03afaed0e6e398e9353d84cbf76835","title":"When NeuroLink uses it","url":"/docs/getting-started/providers/laya#when-neurolink-uses-it","content":"Every built-in consumer of decide — model routing, relevance-driven\ncompaction, tool routing and RAG planning — asks for the default decision\nprovider. That is the first one that is configured — in the environment or in\nthe credentials passed to the SDK — in the order TypeSafe, Laya,\nXOR, SubQ, Perplexity,\nCloudflare Clef, then\nOpenAI Decisions. TypeSafe has two keys,\nTYPESAFE_API_KEY and AI_GATEWAY_API_KEY (its Vercel AI Gateway route), and\neither one counts. Laya counts only with both its key and its base URL, and so\ndoes XOR. SubQ counts with its key alone, SUBQ_API_KEY. Perplexity counts with\nits key alone, and that key,\nPERPLEXITY_API_KEY, is shared with Perplexity's text provider. Cloudflare Clef\ncounts only with both CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID, the same\ntwo variables the Workers AI text provider reads. OpenAI Decisions counts with its\nkey alone, OPENAI_API_KEY, which the OpenAI chat provider reads too. So:\nA TypeSafe key, plus Laya's key and base URL: built-in features use\n  TypeSafe; Laya runs only where a caller asks for provider: \"laya\".\nOnly Laya's key and base URL: built-in features use Laya.\nLaya's key and base URL, plus XOR's: built-in features use Laya; XOR runs\n  only where a caller asks for provider: \"xor\".\nA Laya key with no base URL: Laya is not configured. Built-in features\n  ignore it, and provider: \"laya\" fails with Laya requires a base URL.\nLaya's key and base URL, plus a Perplexity key: built-in features use\n  Laya; Perplexity runs only where a caller asks for\n  provider: \"perplexity-decider\".\nLaya's key and base URL, plus CLOUDFLARE_API_KEY and\n  CLOUDFLARE_ACCOUNT_ID: built-in features use Laya; Cloudflare Clef runs\n  only where a caller asks for provider: \"cloudflare-clef\".\nLaya's key and base URL, plus SUBQ_API_KEY: built-in features use Laya;\n  SubQ runs only where a caller asks for provider: \"subq\".","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"When NeuroLink uses it","lvl3":""}},
+{"objectID":"c99b86a6719e50f022a261b2688e6c849e03afaed0e6e398e9353d84cbf76835","title":"When NeuroLink uses it","url":"/docs/getting-started/providers/laya#when-neurolink-uses-it","content":"Every built-in consumer of decide — model routing, relevance-driven\ncompaction, tool routing and RAG planning — asks for the default decision\nprovider. That is the first one that is configured — in the environment or in\nthe credentials passed to the SDK — in the order TypeSafe, Laya,\nXOR, SubQ, Perplexity,\nCloudflare Clef,\nOpenAI Decisions, then ScaleDown. TypeSafe has two keys,\nTYPESAFE_API_KEY and AI_GATEWAY_API_KEY (its Vercel AI Gateway route), and\neither one counts. Laya counts only with both its key and its base URL, and so\ndoes XOR. SubQ counts with its key alone, SUBQ_API_KEY. Perplexity counts with\nits key alone, and that key,\nPERPLEXITY_API_KEY, is shared with Perplexity's text provider. Cloudflare Clef\ncounts only with both CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID, the same\ntwo variables the Workers AI text provider reads. OpenAI Decisions counts with its\nkey alone, OPENAI_API_KEY, which the OpenAI chat provider reads too.\nScaleDown counts with its key alone, SCALEDOWN_API_KEY, which no other provider\nreads. So:\nA TypeSafe key, plus Laya's key and base URL: built-in features use\n  TypeSafe; Laya runs only where a caller asks for provider: \"laya\".\nOnly Laya's key and base URL: built-in features use Laya.\nLaya's key and base URL, plus XOR's: built-in features use Laya; XOR runs\n  only where a caller asks for provider: \"xor\".\nA Laya key with no base URL: Laya is not configured. Built-in features\n  ignore it, and provider: \"laya\" fails with Laya requires a base URL.\nLaya's key and base URL, plus a Perplexity key: built-in features use\n  Laya; Perplexity runs only where a caller asks for\n  provider: \"perplexity-decider\".\nLaya's key and base URL, plus CLOUDFLARE_API_KEY and\n  CLOUDFLARE_ACCOUNT_ID: built-in features use Laya; Cloudflare Clef runs\n  only where a caller asks for provider: \"cloudflare-clef\".\nLaya's key and base URL, plus SUBQ_API_KEY: built-in features use Laya;\n  SubQ runs only where a caller asks for provider: \"subq\".","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"When NeuroLink uses it","lvl3":""}},
 {"objectID":"adbc69b418b48f0bcaa2ae6e9f87f6bf8efc0dea1fe0d614cc12973156148d9f","title":"Limits","url":"/docs/getting-started/providers/laya#limits","content":"The window is small. Laya's encoders read 1,024 tokens, and 512 on the\nenglish checkpoint, against roughly 33,000 for Jev. Part of that is reserved\nfor each question and its options, so NeuroLink allows about 768 tokens of\nstate on typed-decisions and multilingual, and 320 on english, auto and\nany model name it does not recognise. Laya's server does not refuse a longer\nstate — it answers from the start of it and says nothing — so NeuroLink\nrefuses it instead, before any network call, with max_tokens_exceeded.\n\nThe size is an estimate, not Laya's tokenizer: about four characters per token\nfor ASCII text, and 1.5 tokens per character for other scripts (0.6 on\nmultilingual), calibrated against a live Laya 0.3.5 server. It errs toward refusing. Built-in\nconsumers treat a refusal as \"carry on as before\", which means long-prompt\nmodel routing usually falls back to the heuristic when Laya is the only\ndecision provider.\n\nMany options degrade it. A choice's options share a fixed token budget,\nso accuracy drops past about 20 options; some Laya servers also reject a\nquestion whose options overflow the budget, as invalid_request.\n\nPick the checkpoint deliberately. Laya's own benchmark puts its general\nenglish and multilingual checkpoints close to chance on typed decisions\nwithout fine-tuning; typed-decisions is the fine-tuned one, which is why it is\nthe default.","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"Limits","lvl3":""}},
 {"objectID":"141537feff8c79df3ecbca258b80136e36da0fdeff342aa93a971be020313215","title":"Errors","url":"/docs/getting-started/providers/laya#errors","content":"| Reply                                           | Kind                                                    | Retried |\n| ----------------------------------------------- | ------------------------------------------------------- | ------- |\n| 401 / 403 from the proxy                        | authentication — the provider instance stops retrying | no      |\n| 400 / 422 from Laya                             | invalid_request, with Laya's reason                   | no      |\n| 413 from Laya                                   | max_tokens_exceeded                                   | no      |\n| 429                                             | rate_limit                                            | yes     |\n| 503                                             | overloaded                                            | yes     |\n| other 5xx                                       | server                                                | yes     |\n| more questions or state than Laya reads (local) | max_tokens_exceeded, with no network call             | no      |\n| no base URL configured (local)                  | invalid_request, with no network call                 | no      |\n\nLiteLLM's 401 text echoes a masked copy of the rejected key and its hash; the\nprovider drops that part before the message reaches a log or an error.","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"Errors","lvl3":""}},
 {"objectID":"de358abc64cad163c1666dce7b06fc5f7b3637b9568e420ceb9269a526929d54","title":"Running your own Laya server","url":"/docs/getting-started/providers/laya#running-your-own-laya-server","content":"Laya's example server exposes the same /predict route. Point\nLAYA_BASE_URL at it:","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"Running your own Laya server","lvl3":""}},
 {"objectID":"e0dd67e8f79c7801d3e199dcd765a559f64a9d65d43cf1611026725a8ae1ee82","title":"Troubleshooting","url":"/docs/getting-started/providers/laya#troubleshooting","content":"Laya requires a base URL — set LAYA_BASE_URL, or pass\n  credentials.laya.baseURL. There is no default endpoint.\nAuthentication failed — the proxy rejected the key. Check it in the\n  LiteLLM Dashboard, including its Allowed Routes. The provider instance does\n  not retry after a rejection.\nA 403 with error code: 1010 — Cloudflare in front of your proxy blocked\n  the client by its signature. Node's fetch is accepted; a proxy or agent that\n  rewrites the user agent may not be.\nexceeded the provider's token limit — the state is larger than Laya\n  reads. Shorten it, or configure TypeSafe for long inputs.\nBuilt-in routing never uses Laya — a TypeSafe key is also set and takes\n  precedence, or Laya has no base URL.","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"Troubleshooting","lvl3":""}},
-{"objectID":"fb160b657869a262fd7735599ce2c2ffd49423f4a0f7c43f4ef5ee74aa4dcde5","title":"See also","url":"/docs/getting-started/providers/laya#see-also","content":"The decide inference type\nTypeSafe (Jev) Provider Guide\nXOR Provider Guide\nPerplexity Decisions Provider Guide\nCloudflare Clef Provider Guide\nSubQ Provider Guide\nOpenAI Decisions Provider Guide\nLaya on GitHub","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"See also","lvl3":""}},
+{"objectID":"fb160b657869a262fd7735599ce2c2ffd49423f4a0f7c43f4ef5ee74aa4dcde5","title":"See also","url":"/docs/getting-started/providers/laya#see-also","content":"The decide inference type\nTypeSafe (Jev) Provider Guide\nXOR Provider Guide\nPerplexity Decisions Provider Guide\nCloudflare Clef Provider Guide\nSubQ Provider Guide\nOpenAI Decisions Provider Guide\nScaleDown Provider Guide\nLaya on GitHub","hierarchy":{"lvl0":"Getting Started","lvl1":"Laya Provider Guide","lvl2":"See also","lvl3":""}},
 {"objectID":"df16212357acdf1cff3460cb04e276abe1c702cc845767f49a4bc0cd2a92a763","title":"Lemonfox AI Provider Guide","url":"/docs/getting-started/providers/lemonfox-ai","content":"Lemonfox AI Provider Guide\n\nLemonfox AI is a Tier-2 catalog provider: its integration is one JSON file\n(src/lib/providers/catalog/lemonfox-ai.json) rather than hand-written code.\n\nVerification status: this entry is docs-verified only, not yet\nlive-verified. The vendor's GET https://api.lemonfox.ai/v1/models needs a\nkey (it answers HTTP 401 without one), so the two model ids come from the\nvendor's public chat API page, https://www.lemonfox.ai/apis/chat (retrieved\n2026-09-29), and the roster has not been checked. No account was created and no\nAPI key was used to build it. evidence.liveMatrix is null until someone\nruns the live capability matrix with a real key (see\nVerification status below).\n\nKey Facts\nProvider id: lemonfox-ai (alias lemonfox)\nProtocol: OpenAI-compatible (/chat/completions) — the chat API page says\n  \"Our OpenAI-compatible API takes a list of messages as input and provides an\n  AI-generated (assistant) message as output.\"\nBase URL: https://api.lemonfox.ai/v1\nDefault model: deepseek-v4-flash\nModels in catalog: 2, taken from the model table of\n  https://www.lemonfox.ai/apis/chat\nStreaming: supported — the page's stream parameter reads \"incremental\n  message updates are transmitted as server-sent events with data-only\n  messages.\"\nTool calling: not declared (false) — the page's API Parameters section\n  lists messages, model, max_tokens, stop, stream,\n  frequency_penalty, presence_penalty, temperature and top_p\nTools while streaming: not declared (false)\nStructured output: not declared (false), same parameter list as above\nStructured output + tools together: not declared (false) — no combined\n  probe was possible without credentials\nEmbeddings: not declared\nThinking: not declared\nBilling: free-with-card is the growth queue's classification, one of the\n  three values the catalog schema allows; see Billing below for the\n  vendor's own wording\nKey format: none declared\n\nQuick Start\nGet an API key\nVisit: https://lemonfox.ai/signup (the \"Get started\" and \"Start Your Free Trial\" links on https://www.lemonfox.ai/apis/chat both point to /signup); sign-up was not attempted while building this entry\nCreate an API key at https://lemonfox.ai/apis/keys — https://www.lemonfox.ai/apis/chat links \"create an API key\" to that page, and the API's 401 reply says \"You can obtain an API key from https://lemonfox.ai/apis/keys.\"\nBilling as the vendor states it: see Billing below\nSet LEMONFOX_AI_API_KEY in your .env file\nConfigure\nUse it\n\nPer-request credentials:\n\nBilling\n\nWhat the vendor's public pages show (retrieved 2026-09-29):\nhttps://www.lemonfox.ai/ — \"Try our APIs for 1 month for free.\"\nhttps://www.lemonfox.ai/apis/chat — \"First month for free!\" and \"Start Your\n  Free Trial\"\nhttps://www.lemonfox.ai/terms — Lemon Fox GmbH uses Paddle as its Merchant of\n  Record and offers \"subscription-based pricing and usage-based pricing\"\n\nThe entry records free-with-card, the growth queue's classification, because\nthe catalog schema has no unknown value; it is not a vendor statement of\nsign-up terms.\n\nModels\n\n| Model                  | Context | Price / 1M tokens | Notes                                                                                                                            |\n| ---------------------- | ------- | ----------------- | -------------------------------------------------------------------------------------------------------------------------------- |\n| deepseek-v4-flash ⭐ | 1M      | $0.50             | NeuroLink default. Vendor: \"DeepSeek V4 Flash is our default model — fast, cost-effective, and suitable for most tasks.\"         |\n| mimo-v2.5-pro        | —       | $1.25             | NeuroLink fallback. Vendor: \"MiMo V2.5 Pro is a powerful model from Xiaomi, recommended for complex reasoning and coding tasks.\" |\n\nModel ids, the 1M context figure and the two prices come from the model table\non https://www.lemonfox.ai/apis/chat (retrieved 2026-09-29), whose price column\nis headed \"Price / 1M tokens\". The table's context text for deepseek-v4-flash\nreads \"Supports a 1M token context window.\"; the catalog stores it as 1,000,000\n(NeuroLink's decimal reading of \"1M\"). mimo-v2.5-pro has no context window set.\nThe page also shows the model parameter as \"Specify the model ID to be used.\nDefault: deepseek-v4-flash. Also supported: mimo-v2.5-pro.\"\n\nmodels.defaultContextWindow (32,768) and models.defaultMaxOutputTokens\n(4,096) are placeholders the vendor does not publish, for model ids outside the\ncatalog; the page's max_tokens parameter reads \"integer, optional, default:\ninfinity\". pricingPerMTok is unset on both models because the catalog field\ntakes separate input and output figures and the table has the one \"Price / 1M\ntokens\" column.\n\nEach model's vision flag is false in the catalog, a placeholder for a\nboolean the schema requires; it is not a vendor statement. Each model's\nstatus is production, a value the schema requires.\n\nDeprecated models: a second table on the same page, under the heading\n\"Deprecated","hierarchy":{"lvl0":"Getting Started","lvl1":"Lemonfox AI Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"3fca5359ef393150cf2b043c562fa6d28cdd3d17be9b09b089efaf425beb7ef6","title":"Lemonfox AI Provider Guide","url":"/docs/getting-started/providers/lemonfox-ai#lemonfox-ai-provider-guide","content":"Lemonfox AI is a Tier-2 catalog provider: its integration is one JSON file\n(src/lib/providers/catalog/lemonfox-ai.json) rather than hand-written code.\n\nVerification status: this entry is docs-verified only, not yet\nlive-verified. The vendor's GET https://api.lemonfox.ai/v1/models needs a\nkey (it answers HTTP 401 without one), so the two model ids come from the\nvendor's public chat API page, https://www.lemonfox.ai/apis/chat (retrieved\n2026-09-29), and the roster has not been checked. No account was created and no\nAPI key was used to build it. evidence.liveMatrix is null until someone\nruns the live capability matrix with a real key (see\nVerification status below).","hierarchy":{"lvl0":"Getting Started","lvl1":"Lemonfox AI Provider Guide","lvl2":"Lemonfox AI Provider Guide","lvl3":""}},
 {"objectID":"58c275813c578b02ad4927ca1b96bc8b37a07c75c8829608a8098f8fd53256f1","title":"Key Facts","url":"/docs/getting-started/providers/lemonfox-ai#key-facts","content":"Provider id: lemonfox-ai (alias lemonfox)\nProtocol: OpenAI-compatible (/chat/completions) — the chat API page says\n  \"Our OpenAI-compatible API takes a list of messages as input and provides an\n  AI-generated (assistant) message as output.\"\nBase URL: https://api.lemonfox.ai/v1\nDefault model: deepseek-v4-flash\nModels in catalog: 2, taken from the model table of\n  https://www.lemonfox.ai/apis/chat\nStreaming: supported — the page's stream parameter reads \"incremental\n  message updates are transmitted as server-sent events with data-only\n  messages.\"\nTool calling: not declared (false) — the page's API Parameters section\n  lists messages, model, max_tokens, stop, stream,\n  frequency_penalty, presence_penalty, temperature and top_p\nTools while streaming: not declared (false)\nStructured output: not declared (false), same parameter list as above\nStructured output + tools together: not declared (false) — no combined\n  probe was possible without credentials\nEmbeddings: not declared\nThinking: not declared\nBilling: free-with-card is the growth queue's classification, one of the\n  three values the catalog schema allows; see Billing below for the\n  vendor's own wording\nKey format: none declared","hierarchy":{"lvl0":"Getting Started","lvl1":"Lemonfox AI Provider Guide","lvl2":"Key Facts","lvl3":""}},
@@ -6807,10 +6812,10 @@
 {"objectID":"434f0d25d455ebbe0d3df640f00fe02c0c1b090343b8d702427dae3ab7aa99e7","title":"For local servers, use 'none' or empty string","url":"/docs/getting-started/providers/openai-compatible#for-local-servers-use-none-or-empty-string","content":"OPENAICOMPATIBLEAPI_KEY=none  # ✅ For vLLM\n`","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI-Compatible Providers Guide","lvl2":"For local servers, use 'none' or empty string","lvl3":""}},
 {"objectID":"1de0ce7944bc2cfd6fbf7e218b3d22113f1921563cd445573fb45c834ba477da","title":"Related Documentation","url":"/docs/getting-started/providers/openai-compatible#related-documentation","content":"Provider Setup Guide - General provider configuration\nCost Optimization - Reduce AI costs\nEnterprise Multi-Region - Self-hosted and vLLM deployment","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI-Compatible Providers Guide","lvl2":"Related Documentation","lvl3":""}},
 {"objectID":"dc5208d2366ed5f49efa96638cae619c0fd5aa2eacae848c64a297ac6ec5d07e","title":"Additional Resources","url":"/docs/getting-started/providers/openai-compatible#additional-resources","content":"OpenRouter - Multi-provider aggregator\nvLLM Documentation - Self-hosted inference\nLocalAI - Local OpenAI alternative\nOpenAI API Spec - API standard\n\nNeed Help? Join our GitHub Discussions or open an issue.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI-Compatible Providers Guide","lvl2":"Additional Resources","lvl3":""}},
-{"objectID":"91c74eec73961742b805cee0d280bcf53e9f216a17578ab686dea50f305b0ba4","title":"OpenAI Decisions Provider Guide","url":"/docs/getting-started/providers/openai-decisions","content":"OpenAI Decisions Provider Guide\n\nA provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR,\nPerplexity and Cloudflare Clef,\nfrom OpenAI's Decisions API, and it also reads images. It emits no text at all.\n\nThis is not the OpenAI chat provider (openai), which serves\ngenerate() and stream(). The two share one API key, which has a\nconsequence worth reading before you set it. See\nWhat is sent to OpenAI and\nOne key, two providers.\n\nPublic beta. OpenAI documents the Decisions API as a public beta, so its\nwire may change before general availability. NeuroLink lists its one model as\npreview. Much about the endpoint is not established: see\nWhat is not established. No request in this\nguide, or in NeuroLink's tests, records a reply from the live endpoint; the\nreplies in the tests are stand-ins built from OpenAI's published examples.\n\nOverview\n\ngpt-6-luna is the one model OpenAI's Decisions guide lists for the endpoint.\nYou send one input plus named, typed questions, and optionally images. The\nendpoint returns a typed answer for each question, with a probability instead of\na sentence. NeuroLink calls OpenAI's public endpoint, so a key alone configures\nit: there is no base URL to set.\n\nKey Facts\n\n|                       |                                                                                                                                                                                                                                                                                                                                                                                           |\n| --------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Provider id           | openai-decisions (no aliases) — distinct from openai, the chat provider                                                                                                                                                                                                                                                                                                               |\n| Inference type        | decide only                                                                                                                                                                                                                                                                                                                                                                             |\n| Model                 | gpt-6-luna; OPENAI_DECISIONS_MODEL or a per-call model sets it                                                                                                                                                                                                                                                                                                                      |\n| Key                   | OPENAI_API_KEY, shared with the OpenAI chat provider, or credentials.openaiDecisions.apiKey                                                                                                                                                                                                                                                                                           |\n| Endpoint              | https://api.openai.com/v1/decisions; OPENAI_DECISIONS_BASE_URL can name another origin                                                                                                                                                                                                                                                                                                |\n| Media                 | images as inline data URLs, up to 8 per request (NeuroLink's own allowance, not OpenAI's); the endpoint has no video part                                                                                                                                                                                                                                                                 |\n| Request body cap      | 32 MiB on the encoded request, NeuroLink's own allowance, not OpenAI's                                                                                                                                                                                                                                                                                                                    |\n| State window          | 922,000 tokens is reported as an advisory figure (the model page's input figure); NeuroLink neve","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"","lvl3":""}},
+{"objectID":"91c74eec73961742b805cee0d280bcf53e9f216a17578ab686dea50f305b0ba4","title":"OpenAI Decisions Provider Guide","url":"/docs/getting-started/providers/openai-decisions","content":"OpenAI Decisions Provider Guide\n\nA provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR,\nPerplexity and Cloudflare Clef,\nfrom OpenAI's Decisions API, and it also reads images. It emits no text at all.\n\nThis is not the OpenAI chat provider (openai), which serves\ngenerate() and stream(). The two share one API key, which has a\nconsequence worth reading before you set it. See\nWhat is sent to OpenAI and\nOne key, two providers.\n\nPublic beta. OpenAI documents the Decisions API as a public beta, so its\nwire may change before general availability. NeuroLink lists its one model as\npreview. Much about the endpoint is not established: see\nWhat is not established. No request in this\nguide, or in NeuroLink's tests, records a reply from the live endpoint; the\nreplies in the tests are stand-ins built from OpenAI's published examples.\n\nOverview\n\ngpt-6-luna is the one model OpenAI's Decisions guide lists for the endpoint.\nYou send one input plus named, typed questions, and optionally images. The\nendpoint returns a typed answer for each question, with a probability instead of\na sentence. NeuroLink calls OpenAI's public endpoint, so a key alone configures\nit: there is no base URL to set.\n\nKey Facts\n\n|                       |                                                                                                                                                                                                                                                                                                                                                                                                                                         |\n| --------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Provider id           | openai-decisions (no aliases) — distinct from openai, the chat provider                                                                                                                                                                                                                                                                                                                                                             |\n| Inference type        | decide only                                                                                                                                                                                                                                                                                                                                                                                                                           |\n| Model                 | gpt-6-luna; OPENAI_DECISIONS_MODEL or a per-call model sets it                                                                                                                                                                                                                                                                                                                                                                    |\n| Key                   | OPENAI_API_KEY, shared with the OpenAI chat provider, or credentials.openaiDecisions.apiKey                                                                                                                                                                                                                                                                                                                                         |\n| Endpoint              | https://api.openai.com/v1/decisions; OPENAI_DECISIONS_BASE_URL can name another origin                                                                                                                                                                                                                                                                                                                                              |\n| Media                 | images as inline data URLs, up to 8 per request (NeuroLink's own allowance, not OpenAI's); the endpoint has no video part                                                                                                                                                                                                                                                                                                               |\n| Request body cap      | 32 MiB on the encoded request, NeuroLink's own allowance, not OpenAI's                                                                ","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"e277b3d342fd3fb79433f472e660f6c52f72efe27989a4d67d5484fd148cc161","title":"OpenAI Decisions Provider Guide","url":"/docs/getting-started/providers/openai-decisions#openai-decisions-provider-guide","content":"A provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR,\nPerplexity and Cloudflare Clef,\nfrom OpenAI's Decisions API, and it also reads images. It emits no text at all.\n\nThis is not the OpenAI chat provider (openai), which serves\ngenerate() and stream(). The two share one API key, which has a\nconsequence worth reading before you set it. See\nWhat is sent to OpenAI and\nOne key, two providers.\n\nPublic beta. OpenAI documents the Decisions API as a public beta, so its\nwire may change before general availability. NeuroLink lists its one model as\npreview. Much about the endpoint is not established: see\nWhat is not established. No request in this\nguide, or in NeuroLink's tests, records a reply from the live endpoint; the\nreplies in the tests are stand-ins built from OpenAI's published examples.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"OpenAI Decisions Provider Guide","lvl3":""}},
 {"objectID":"49eb266610612a7852f59caabbf8ad2c11b45f0cf611089a69ea264bc34a4851","title":"Overview","url":"/docs/getting-started/providers/openai-decisions#overview","content":"gpt-6-luna is the one model OpenAI's Decisions guide lists for the endpoint.\nYou send one input plus named, typed questions, and optionally images. The\nendpoint returns a typed answer for each question, with a probability instead of\na sentence. NeuroLink calls OpenAI's public endpoint, so a key alone configures\nit: there is no base URL to set.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"Overview","lvl3":""}},
-{"objectID":"bf5a6a94239f41122a6e62abc5358dd4226541e57007bc5b220a28229b86fbc9","title":"Key Facts","url":"/docs/getting-started/providers/openai-decisions#key-facts","content":"|                       |                                                                                                                                                                                                                                                                                                                                                                                           |\n| --------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Provider id           | openai-decisions (no aliases) — distinct from openai, the chat provider                                                                                                                                                                                                                                                                                                               |\n| Inference type        | decide only                                                                                                                                                                                                                                                                                                                                                                             |\n| Model                 | gpt-6-luna; OPENAI_DECISIONS_MODEL or a per-call model sets it","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"Key Facts","lvl3":""}},
+{"objectID":"bf5a6a94239f41122a6e62abc5358dd4226541e57007bc5b220a28229b86fbc9","title":"Key Facts","url":"/docs/getting-started/providers/openai-decisions#key-facts","content":"|                       |                                                                                                                                                                                                                                                                                                                                                                                                                                         |\n| --------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Provider id           | openai-decisions (no aliases) — distinct from openai, the chat provider                                                                                                                                                                                                                                                                                                                                                             |\n| Inference type        | decide only                                                                                                                                                                                                                                                                                                                                                                                                                           |\n| Model                 | gpt-6-luna; OPENAI_DECISIONS_MODEL or a per-call model sets it","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"Key Facts","lvl3":""}},
 {"objectID":"399bde759d649245ede8e1d2215c73851d6446e43dfeb6c5dabf5d715353ed19","title":"1. Get a key","url":"/docs/getting-started/providers/openai-decisions#1-get-a-key","content":"Create one in the OpenAI dashboard. It is\nthe same kind of key the OpenAI chat provider uses.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"1. Get a key","lvl3":""}},
 {"objectID":"08b1702a4ad8240cce1dc774b5326214664f7c6fae91197a0dff51d62cce7155","title":"2. Configure","url":"/docs/getting-started/providers/openai-decisions#2-configure","content":"Set the key in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:\n\ncredentials.openaiDecisions is its own slice (apiKey and baseURL). A key\npassed as credentials.openai belongs to the chat provider and does not\nconfigure this one. OPENAI_BASE_URL, which moves the chat provider, never\nroutes a decision, and neither does OPENAI_MODEL pick its model; the overrides\nfor this provider are OPENAI_DECISIONS_BASE_URL and OPENAI_DECISIONS_MODEL.\nOPENAI_BASE_URL, OPENAI_API_BASE and the SDK's credentials.openai.baseURL\nare read for one purpose only: to tell whether the environment key is OpenAI's\nown, so that a gateway's key is not used for the default; see\nA gateway's key.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"2. Configure","lvl3":""}},
 {"objectID":"d1990335200138e7298ff0afc1ecb603368cb5543cb1bfe5450e78ac7f2f6763","title":"3. Use it","url":"/docs/getting-started/providers/openai-decisions#3-use-it","content":"A boolean answer carries a probability and no confidence of its own. A choice\nor score answer carries the confidence the endpoint reports; when a reply\nomits it, NeuroLink uses the highest probability in that answer's distribution\ninstead. Whether the endpoint's confidence is calibrated is not established, so\ntune a threshold on it against your own labelled data.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"3. Use it","lvl3":""}},
@@ -6819,16 +6824,16 @@
 {"objectID":"df550a22381e831fbd8e7b17e43ed09b0b2c49d33ef518c609867ed8d72c09af","title":"Worked example","url":"/docs/getting-started/providers/openai-decisions#worked-example","content":"These are the three examples from\nOpenAI's Decisions guide\n(saved 2026-10-07), as NeuroLink sends and reads them. The questions:\n\ngo out as:\n\nThe guide's example answers, read back by name:\n\n| Question         | Answer in the guide                                                                       | result.answers[name]                                                                                                                       |\n| ---------------- | ----------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------- |\n| visible_damage | probability 0.92                                                                        | { type: \"boolean\", probability: 0.92 }                                                                                                     |\n| department     | choice billing; probabilities 0.95 / 0.02 / 0.01 / 0.02; confidence 0.93            | { type: \"choice\", choice: \"billing\", confidence: 0.93, probabilities: { billing: 0.95, technical: 0.02, shipping: 0.01, other: 0.02 } }    |\n| severity       | score 1.1; probabilities 0.1 / 0.7 / 0.2 for levels 0, 1 and 2; confidence 0.55 | { type: \"score\", score: 1.1, confidence: 0.55, probabilities: { \"0\": 0.1, \"1\": 0.7, \"2\": 0.2 }, legend: { \"0\": \"Appearance only; …\", … } } |\n\nThe score is the probability-weighted average of the level indices (0 × 0.1 +\n1 × 0.7 + 2 × 0.2 = 1.1), the same meaning as NeuroLink's score everywhere. The\nendpoint returns no legend, so NeuroLink builds legend from the rubric you\nsent, keyed by level index, and a level with no text has the empty string. A\nlevel label in the reply is not read: the integer value identifies the level.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"Worked example","lvl3":""}},
 {"objectID":"e32095e540af0c39d6aae4a574d82e9d3cc00d55996e76dd1e88d5334fde4b58","title":"What a reply must look like","url":"/docs/getting-started/providers/openai-decisions#what-a-reply-must-look-like","content":"NeuroLink checks each answer against the question it asked and drops an answer\nthat fails any check, with a warning that names the question and never quotes the\nreply. It does not repair or default an answer, and the other answers in the same\nreply survive. An answer is dropped when:\nit has no name, or a name NeuroLink did not ask, or the same name appears\n  more than once (every answer under a repeated name is dropped);\nits type is not the type asked (predicate, choice or score);\na probability is not a number from 0 to 1;\na choice's choice is not one of the options asked, or its probabilities do\n  not list each asked option exactly once;\na score's probabilities do not list each level index exactly once as an\n  integer value, or its score is not a number from 0 to the highest index;\nits confidence is a number outside 0 to 1 (a missing or non-numeric one is\n  replaced by the highest probability in the distribution);\nits probabilities do not sum to 1, within a tolerance of 0.01 (with a rounding\n  slack of 1e-9 for floating point). Whether OpenAI rounds the probabilities it\n  returns, and by how much, is not established; an answer whose rounding drifts\n  past the tolerance is dropped like any other.\n\nA refusal answer (type: \"refusal\") means the model declined that question.\nNeuroLink has no refusal answer type: the question is simply absent from\nresult.answers, and a warning records the model declined to answer \"<id>\".\nA question the reply never mentions is absent too, without a warning. Every\ninternal consumer treats an absent answer as no decision. A reply whose answers\nis not an array fails as a server error.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"What a reply must look like","lvl3":""}},
 {"objectID":"b4619d81305066595c1fe14d0bd051d2a4534341b991a72c7b6ca8172d877bc7","title":"Images","url":"/docs/getting-started/providers/openai-decisions#images","content":"OpenAI's Decisions API reads images, and no video part exists. Pass them as for\nany decide provider that reads images:\n\nor neurolink decide ... --provider openai-decisions --image ./photo.png from the\nCLI. An image can be a Buffer, a file path or a data: URL; a remote http(s)\nURL is not fetched for you. Every image is sent inline as a base64 data URL in an\ninput_image part. The endpoint takes no hosted URL and no file id, and\nNeuroLink sets no detail.\n\nNeuroLink refuses these before any request, as a non-retried invalid_request:\nmore than 8 images, which is NeuroLink's own allowance (the same number it\n  allows XOR and Perplexity), not a limit OpenAI documents;\na video, because the endpoint has no video part;\nan image NeuroLink cannot recognise from its bytes, a file that cannot be read,\n  or an encoded request body over 32 MiB (also NeuroLink's own number).\n\nThe accepted image formats, the per-request image limit and the body cap on\nOpenAI's side are not established. NeuroLink does not check the format against a\nlist of OpenAI's, because it has none to check against.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"Images","lvl3":""}},
-{"objectID":"9ea430557ccbea3760eeb08f9ddfc727349d78df4612512eb4c634fc537d9e93","title":"When NeuroLink uses it","url":"/docs/getting-started/providers/openai-decisions#when-neurolink-uses-it","content":"Every built-in consumer of decide asks for the default decision provider. That\nis the first one that is configured, in the environment or in the credentials\npassed to the SDK, in the order TypeSafe, Laya, XOR, SubQ, Perplexity,\nCloudflare Clef, then OpenAI Decisions. TypeSafe counts\nwith either of its keys, TYPESAFE_API_KEY or AI_GATEWAY_API_KEY. Laya and XOR\neach count only with both their key and their base URL. Perplexity counts with\nits key alone. Cloudflare Clef counts only with both CLOUDFLARE_API_KEY and\nCLOUDFLARE_ACCOUNT_ID. OpenAI Decisions counts with its key alone, a non-blank\nOPENAI_API_KEY or credentials.openaiDecisions.apiKey, except that a key found\nonly in the environment does not count while OPENAI_BASE_URL,\nOPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com: see\nA gateway's key is never used for the default.\nA caller can always name it with provider: \"openai-decisions\". So:\nOnly an OpenAI key, and no base URL for another host: built-in features use\n  OpenAI Decisions.\nAn OpenAI key beside OPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL naming a gateway:\n  the key is taken to be the gateway's, and built-in features behave exactly as\n  they did without a decision model, unless OPENAI_DECISIONS_BASE_URL is set or\n  a credentials.openaiDecisions key or base URL is passed.\nAn OpenAI key, plus any key or credentials that configure an earlier provider:\n  built-in features use the earlier provider, in the order above. OpenAI\n  Decisions runs only where a caller asks for provider: \"openai-decisions\".\nA key for the chat provider only: the same as the first case. The key is\n  shared, so the chat provider's key also configures this one. See\n  One key, two providers.\nNone of TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef or OpenAI Decisions\n  configured: everything behaves exactly as it did without a decision model.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"When NeuroLink uses it","lvl3":""}},
+{"objectID":"9ea430557ccbea3760eeb08f9ddfc727349d78df4612512eb4c634fc537d9e93","title":"When NeuroLink uses it","url":"/docs/getting-started/providers/openai-decisions#when-neurolink-uses-it","content":"Every built-in consumer of decide asks for the default decision provider. That\nis the first one that is configured, in the environment or in the credentials\npassed to the SDK, in the order TypeSafe, Laya, XOR, SubQ, Perplexity,\nCloudflare Clef, OpenAI Decisions, then\nScaleDown. TypeSafe counts\nwith either of its keys, TYPESAFE_API_KEY or AI_GATEWAY_API_KEY. Laya and XOR\neach count only with both their key and their base URL. Perplexity counts with\nits key alone. Cloudflare Clef counts only with both CLOUDFLARE_API_KEY and\nCLOUDFLARE_ACCOUNT_ID. OpenAI Decisions counts with its key alone, a non-blank\nOPENAI_API_KEY or credentials.openaiDecisions.apiKey, except that a key found\nonly in the environment does not count while OPENAI_BASE_URL,\nOPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com: see\nA gateway's key is never used for the default.\nScaleDown counts with its key alone, SCALEDOWN_API_KEY, which no other provider\nreads. A caller can always name it with provider: \"openai-decisions\". So:\nOnly an OpenAI key, and no base URL for another host: built-in features use\n  OpenAI Decisions.\nAn OpenAI key beside OPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL naming a gateway:\n  the key is taken to be the gateway's, and built-in features behave exactly as\n  they did without a decision model, unless OPENAI_DECISIONS_BASE_URL is set or\n  a credentials.openaiDecisions key or base URL is passed.\nAn OpenAI key, plus any key or credentials that configure an earlier provider:\n  built-in features use the earlier provider, in the order above. OpenAI\n  Decisions runs only where a caller asks for provider: \"openai-decisions\".\nAn OpenAI key, plus SCALEDOWN_API_KEY: built-in features use OpenAI\n  Decisions, which comes ahead of ScaleDown. ScaleDown runs only where a caller\n  asks for provider: \"scaledown\".\nA key for the chat provider only: the same as the first case.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"When NeuroLink uses it","lvl3":""}},
 {"objectID":"4c0c98775b0c21bdaeffbd51a77736e61ca1847dfea62204eab148c02317a84d","title":"What is sent to OpenAI","url":"/docs/getting-started/providers/openai-decisions#what-is-sent-to-openai","content":"OpenAI receives the state and the questions of every decision made on your\nbehalf. The questions are NeuroLink's fixed wording, plus text from your own\nconversation where a consumer quotes it; the state is where your text goes.\nThis is what each built-in consumer sends and what turns it on, read from the\ncode:\nModel routing. One request behind the router, the model catalogue and the\n  per-request context budget.\nSends: the first 8,000 characters of the prompt of the call being routed,\n    plus whether tools are available, whether the request includes images, the\n    estimated input tokens, the requested thinking level, whether a session is\n    bound and how many prior messages the caller passed in. When the pool has\n    more than one model, the questions carry the ids and descriptions of the\n    models in it. The images, the session id and the conversation itself are not\n    sent.\nTurned on by: classifierRouter.enabled with a pool: an opt-in. With\n    the default classifier: \"auto\", a configured decision provider selects the\n    decision strategy. classifier: \"heuristic\" or \"llm\" never calls a\n    decision provider. A call that pins both provider and model is not\n    routed.\nRelevance compaction. Stage 0 of context compaction.\nSends: the current request (first 4,000 characters) and, for each earlier\n    user or assistant message that is plain text (not a tool call or result, a\n    summary or a pinned skill), its role and its first 1,200 characters. That\n    text travels twice, once in the state and once in the message's question. By\n    default the six most recent messages are never sent\n    (contextRelevance.protectRecent), and at most 300 are asked about.\nTurned on by: nothing beyond a configured decision provider. It runs\n    inside context compaction, which starts when a conversation outgrows its\n    context budget.\nSummary-quality gate. Guards the summarization stage of compaction.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"What is sent to OpenAI","lvl3":""}},
-{"objectID":"ebe4c10fa9c8ba932d0fa55fd540a3c022964f71e344c17bc1f01673aa141398","title":"One key, two providers","url":"/docs/getting-started/providers/openai-decisions#one-key-two-providers","content":"OPENAI_API_KEY is the variable the OpenAI chat provider reads to\nserve generate() and stream(). It is also the variable that configures this\nprovider, and nothing else needs to be set. This changes what a host with only\nan OpenAI key does by default. A host that set the key only to use OpenAI chat\nhas also configured decide:\nWith none of TypeSafe, Laya, XOR, SubQ, Perplexity or Cloudflare Clef configured,\n  the texts listed in What is sent to OpenAI go to\n  OpenAI's Decisions API, and each decision is a paid call. Context compaction\n  needs no other opt-in, so a host with long conversations starts sending\n  conversation text the first time one outgrows its budget. Model routing, tool\n  routing and RAG planning do so only where you have turned them on. A host that\n  passes a transcription dictionary also sends transcript text. No code changes:\n  the key is the switch.\nIt is fail-open, and every other decision provider takes precedence. If you\n  have configured any of them, the shared key changes nothing for the built-in\n  consumers. A failure here leaves behaviour as it was, so a key the Decisions\n  endpoint does not accept costs requests, not correctness.\nA rejected key is not remembered across calls. A 401 stops one provider\n  instance from retrying, but neurolink.decide() and neurolink.tryDecide()\n  build a fresh provider for every call, so the breaker never latches across\n  them. A host whose key the endpoint rejects makes one failing request for each\n  decision site that runs, every time it runs.\nWith no decision provider configured at all, nothing is sent.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"One key, two providers","lvl3":""}},
+{"objectID":"ebe4c10fa9c8ba932d0fa55fd540a3c022964f71e344c17bc1f01673aa141398","title":"One key, two providers","url":"/docs/getting-started/providers/openai-decisions#one-key-two-providers","content":"OPENAI_API_KEY is the variable the OpenAI chat provider reads to\nserve generate() and stream(). It is also the variable that configures this\nprovider, and nothing else needs to be set. This changes what a host with only\nan OpenAI key does by default. A host that set the key only to use OpenAI chat\nhas also configured decide:\nWith none of TypeSafe, Laya, XOR, SubQ, Perplexity or Cloudflare Clef configured,\n  the texts listed in What is sent to OpenAI go to\n  OpenAI's Decisions API, and each decision is a paid call. Context compaction\n  needs no other opt-in, so a host with long conversations starts sending\n  conversation text the first time one outgrows its budget. Model routing, tool\n  routing and RAG planning do so only where you have turned them on. A host that\n  passes a transcription dictionary also sends transcript text. No code changes:\n  the key is the switch.\nIt is fail-open, and the six providers ahead of it take precedence. If you\n  have configured any of them, the shared key changes nothing for the built-in\n  consumers. A failure here leaves behaviour as it was, so a key the Decisions\n  endpoint does not accept costs requests, not correctness.\nA rejected key is not remembered across calls. A 401 stops one provider\n  instance from retrying, but neurolink.decide() and neurolink.tryDecide()\n  build a fresh provider for every call, so the breaker never latches across\n  them. A host whose key the endpoint rejects makes one failing request for each\n  decision site that runs, every time it runs.\nWith no decision provider configured at all, nothing is sent.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"One key, two providers","lvl3":""}},
 {"objectID":"589d4b33834cecbfd79e7a3eeac0c17d5ba4b2b1f72e19c703f9495f7604207d","title":"A gateway's key is never used for the default","url":"/docs/getting-started/providers/openai-decisions#a-gateways-key-is-never-used-for-the-default","content":"OPENAI_API_KEY is also the key of every OpenAI-compatible gateway (LiteLLM,\nvLLM, LM Studio and the like), and a host that points the chat provider at one\nsets OPENAI_BASE_URL (or, in LangChain-style setups, OPENAI_API_BASE), or an SDK\nhost passes credentials.openai.baseURL, to it. The chat provider prefers the\ncredentials field over the variable.\nThat key belongs to the gateway, and the decision text belongs on the gateway's\nside of the wire, so NeuroLink does not take such a key to be OpenAI's: while\nOPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL names a host other than\nhttps://api.openai.com, the environment key is not used to pick OpenAI Decisions\nas the default. The built-in consumers send nothing to api.openai.com for it:\neach behaves as it did without a decision model, and the \"no decision provider\"\nerror says why the key is ignored. Only a call that names the provider reaches\nOpenAI, as described below.\n\nThe check reads each value's origin, that is scheme, host and port, compared\nwithout regard to case. A path, a trailing slash and /v1 do not matter.\n\n| OPENAI_BASE_URL / OPENAI_API_BASE / credentials.openai.baseURL                | The environment key is used for the default |\n| ----------------------------------------------------------------------------------- | ------------------------------------------- |\n| unset, empty or blank                                                               | yes                                         |\n| https://api.openai.com, https://api.openai.com/v1/, https://API.openai.com    | yes                                         |\n| https://gateway.example/v1, http://127.0.0.1:4000/v1, http://localhost:11434  | no                                          |\n| https://eu.api.openai.com, http://api.openai.com, https://api.openai.com:8443 | no: another host, scheme or port            |\n| api.openai.com (no scheme), not a url, ftp://...","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"A gateway's key is never used for the default","lvl3":""}},
 {"objectID":"4578834b017bbed1b8de1a71d677386e28437557be3814e141fa7becf3e36be2","title":"What is not established","url":"/docs/getting-started/providers/openai-decisions#what-is-not-established","content":"Each item below is not measured: nothing NeuroLink was built or tested from\nsays what the endpoint does, and no live request was made. Where NeuroLink had\nto choose, it chose to send only documented fields, to refuse nothing locally on\nOpenAI's behalf, and to drop a bad answer rather than guess.\nWhether /v1/decisions uses OpenAI's general error envelope, and the status\n   and text of a refusal for an over-length input. NeuroLink reads an error's\n   error.message when there is one, and classifies by HTTP status.\nThe maximum number of questions per request, of choices per question (the\n   benchmarks NeuroLink has in mind need 77 and 151), and of levels per score,\n   and what a question name may contain.\nThe endpoint's input ceiling (the model page gives 922,000 input tokens for the\n   model in general), the number of images per request, the request body cap and\n   the accepted image formats.\nWhether probabilities always covers every supplied option, in what order, and\n   how the numbers are rounded. NeuroLink reads each distribution by key, not by\n   position, and requires each asked option exactly once.\nThe endpoint's rate limits. The model page lists limits for the model, not for\n   this endpoint.\nHow long a request takes, for any size. The 10-second default timeout is a\n   starting point.\nWhether a regional host such as https://eu.api.openai.com serves this\n   endpoint with the same request. The base URL can be changed to try one.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"What is not established","lvl3":""}},
 {"objectID":"155654a19ca65d0e67082bb342d58723138287406b9783055a27d57d44812646","title":"Cost","url":"/docs/getting-started/providers/openai-decisions#cost","content":"$0.10 per 1M input tokens, input only, from\nOpenAI's Decisions guide\nas read on 2026-10-07. Regional and long-context multipliers are not modelled.\nNeuroLink prices a decision from the input tokens the reply reports and counts\noutput tokens at nothing, so result.usage and the CLI's Cost: line are an\nestimate from that one rate, not a bill. A reply that reports no usage is priced\nat zero.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"Cost","lvl3":""}},
 {"objectID":"f61611df66cdbcf4e2fb52b5095b4ec859a6a96289fc4e4be0a383f8484f3c9f","title":"Errors","url":"/docs/getting-started/providers/openai-decisions#errors","content":"| Reply                                                                | Kind                                                    | Retried                                  |\n| -------------------------------------------------------------------- | ------------------------------------------------------- | ---------------------------------------- |\n| 401                                                                  | authentication — the provider instance stops retrying | no                                       |\n| 429 (every one)                                                      | rate_limit                                            | no                                       |\n| 503                                                                  | overloaded                                            | yes, once (after Retry-After, if sent) |\n| other 5xx (500, 502, 504)                                            | server                                                | yes, once                                |\n| any other status, including 400, 402, 403 and 404                    | invalid_request, with the reply's own message         | no                                       |\n| a 200 whose answers is not an array                                | server                                                | no                                       |\n| no response within the timeout                                       | timeout                                               | yes, once                                |\n| the request fails before any reply (connection refused, DNS failure) | network                                               | yes, once                                |\n| the caller's own signal aborts the request                         | network                                               | no                                       |\n| more than 8 images, a video, or an unusable image (local)            | invalid_request, with no network call","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"Errors","lvl3":""}},
 {"objectID":"7cdf00f2dd7ac7f9cfae6a1b387567d1cf0c7de8a98f5c47210a6ff8221db4d6","title":"Base URL","url":"/docs/getting-started/providers/openai-decisions#base-url","content":"OPENAI_DECISIONS_BASE_URL, or credentials.openaiDecisions.baseURL, replaces\nhttps://api.openai.com. It is an origin:\ntrailing slashes and one trailing /v1 are stripped, so\n  https://api.openai.com/v1/ and https://api.openai.com are the same;\nhttp:// is accepted only for localhost, 127.0.0.1 and [::1], because the\n  request carries the key and the state. Any other host needs https://; 0.0.0.0\n  and the rest of 127.0.0.0/8 do not count as this machine;\na user name or password, a query string, a fragment, a scheme other than\n  http or https, http:// to a host other than those three, or a value that\n  is not an absolute URL is refused before any request, with a message that never\n  repeats the value;\na redirect from the endpoint is refused rather than followed, so the decision is\n  never replayed at an address the server chooses;\nthe request goes to <origin>/v1/decisions.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"Base URL","lvl3":""}},
 {"objectID":"c6c6f00939ee73c7416d6eed5fcb54dfca58a77c31a358c0e1b630c25fdf9e71","title":"Troubleshooting","url":"/docs/getting-started/providers/openai-decisions#troubleshooting","content":"No decision provider is configured although OPENAI_API_KEY is set —\n  OPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL names a host other than\n  https://api.openai.com, so the key is taken to be a gateway's. See\n  A gateway's key: set\n  OPENAI_DECISIONS_BASE_URL or pass credentials.openaiDecisions to route this\n  provider on purpose. Naming the provider in a call is not a way out when the key\n  is a gateway's: a named call sends the environment key to api.openai.com.\nOpenAI Decisions requires an API key — set OPENAI_API_KEY, or pass\n  credentials.openaiDecisions.apiKey. A key passed as credentials.openai does\n  not count.\nA 401 — OpenAI rejected the key. The provider instance stops retrying after\n  a rejection, so fix the key and construct a new one.\nA 403 — OpenAI's general error documentation lists a 403 for an unsupported\n  country; what this endpoint's own 403s mean is not established. It does not latch\n  the instance, so a change on the account side works at once.\nA 400 naming the model — OPENAI_DECISIONS_MODEL, or a per-call model,\n  names something other than a model the endpoint supports.\nThe OpenAI Decisions base URL must not carry credentials… — the override\n  has a user name, a password, a query string or a fragment. Set it to the origin\n  only, or leave it unset for https://api.openai.com.\nThe OpenAI Decisions base URL must start with https:// or http:// — the\n  value parses as a URL with another scheme, or is a host name and port with no\n  scheme, such as localhost:8080, which reads as the scheme localhost:. Put\n  https:// in front, or http:// for a server on this machine.\nThe OpenAI Decisions base URL must use https:// unless it points at this machine —\n  the override is an http:// URL whose host is not localhost, 127.0.0.1 or\n  [::1]. A plain-http request would send the key and the state unencrypted. Use\n  https://, or reach the server through a port-forward to localhost.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"Troubleshooting","lvl3":""}},
-{"objectID":"b7b92543296a3f4ebd2f58c8d0c68da7750bb2a51df7eb7abaca356794a4681b","title":"See also","url":"/docs/getting-started/providers/openai-decisions#see-also","content":"The decide inference type\nTypeSafe (Jev) Provider Guide\nLaya Provider Guide\nXOR Provider Guide\nPerplexity Decisions Provider Guide\nCloudflare Clef Provider Guide\nOpenAI chat provider\nOpenAI's Decisions guide","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"See also","lvl3":""}},
+{"objectID":"b7b92543296a3f4ebd2f58c8d0c68da7750bb2a51df7eb7abaca356794a4681b","title":"See also","url":"/docs/getting-started/providers/openai-decisions#see-also","content":"The decide inference type\nTypeSafe (Jev) Provider Guide\nLaya Provider Guide\nXOR Provider Guide\nPerplexity Decisions Provider Guide\nCloudflare Clef Provider Guide\nScaleDown Provider Guide\nOpenAI chat provider\nOpenAI's Decisions guide","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Decisions Provider Guide","lvl2":"See also","lvl3":""}},
 {"objectID":"d6669edf9f3a9ed44443cdbca83140ef76b3dffb6c23994cb07ca53b8661aece","title":"OpenAI TTS Provider Guide","url":"/docs/getting-started/providers/openai-tts","content":"OpenAI TTS Provider Guide\n\nHigh-quality neural text-to-speech with six distinct voices and HD quality option\n\nOverview\n\nNeuroLink integrates OpenAI's Text-to-Speech API, giving you access to six expressive neural voices across two model tiers. The standard model (tts-1) optimises for low latency, while the HD model (tts-1-hd) delivers higher audio fidelity for production use cases such as podcasts, voice assistants, and narration.\n\nOpenAI TTS works with any NeuroLink text generation call — you can synthesise the raw prompt directly or synthesise the AI-generated response, controlled by the useAiResponse flag.\n\nKey Facts\n\n| Property         | Value                                         |\n| ---------------- | --------------------------------------------- |\n| Provider ID  | openai-tts                                  |\n| API endpoint | https://api.openai.com/v1/audio/speech      |\n| Models       | tts-1 (standard), tts-1-hd (high quality) |\n| Voices       | alloy, echo, fable, onyx, nova, shimmer       |\n| Formats      | mp3, wav, ogg (opus), opus                    |\n| Max input    | 4,096 characters per request                  |\n| Languages    | Follows input text language automatically     |\n| Streaming    | Not supported (batch synthesis only)          |\n\nQuick Start\nGet an API Key\n\nSign up or log in at https://platform.openai.com and create a new secret key under API keys.\nConfigure Environment\n\nAdd to your .env file:\n\nDefault model (tts-1) and voice (alloy) are set in code; override per call\nvia tts.model / tts.voice in the SDK or --tts-model / --tts-voice on\nthe CLI. There is no OPENAI_TTS_MODEL / OPENAI_TTS_VOICE env var.\nInstall NeuroLink\nSynthesise Your First Audio\n\nSupported Models\n\n| Model ID   | Quality  | Latency | Use Case                                       |\n| ---------- | -------- | ------- | ---------------------------------------------- |\n| tts-1    | Standard | Lower   | Default; real-time apps, interactive voice UIs |\n| tts-1-hd | HD       | Higher  | Podcasts, narration, production audio assets   |\n\nSelect the HD model by passing quality: \"hd\" in TTS options — NeuroLink maps this automatically to tts-1-hd.\n\nSDK Usage\n\nDirect Text Synthesis\n\nSynthesise the input text directly without calling an AI model:\n\nAI Response Synthesis\n\nGenerate a response with an AI model and then synthesise it:\n\nHD Quality Audio\n\nAdjusting Playback Speed\n\nSave to File\n\nPer-Call Credential Override\n\nCLI Usage\n\nBasic TTS\n\nHD Quality\n\nSynthesise AI Response\n\nSpeed Adjustment\n\nAvailable Voices\n\n| Voice ID  | Gender  | Character                        | Best For                          |\n| --------- | ------- | -------------------------------- | --------------------------------- |\n| alloy   | Neutral | Balanced, clear, versatile       | General purpose, default          |\n| echo    | Male    | Crisp, authoritative             | Announcements, business content   |\n| fable   | Neutral | Warm, expressive, storytelling   | Narration, audiobooks             |\n| onyx    | Male    | Deep, confident, professional    | Voiceovers, documentary           |\n| nova    | Female  | Bright, friendly, conversational | Voice assistants, customer-facing |\n| shimmer | Female  | Soft, gentle, calm               | Wellness apps, guided meditation  |\n\nOpenAI voices are language-agnostic — they follow the language of the input text automatically, supporting English, Spanish, French, German, Japanese, and many more.\n\nAudio Formats\n\n| Format | Extension | Use Case                                | Notes                |\n| ------ | --------- | --------------------------------------- | -------------------- |\n| mp3  | .mp3    | Default; web, mobile, general storage   | 24 kHz sample rate   |\n| wav  | .wav    | Uncompressed; audio editors, processing | 24 kHz sample rate   |\n| ogg  | .ogg    | Browser streaming, web apps             | Opus codec at 48 kHz |\n| opus | .opus   | Low-bandwidth streaming                 | Opus codec at 48 kHz |\n\nConfiguration Reference\n\n| Environment Variable | Required | Default | Description                        |\n| -------------------- | -------- | ------- | ---------------------------------- |\n| OPENAI_API_KEY     | Yes      | —       | OpenAI API key (starts with sk-) |\n\nDefaults for model (tts-1) and voice (alloy) are set in code; override per\ncall via tts.model / tts.voice in the SDK or --tts-model / --tts-voice\non the CLI. There are no OPENAI_TTS_MODEL / OPENAI_TTS_VOICE env vars.\n\nFeature Support Matrix\n\n| Feature                | Supported | Notes                              |\n| ---------------------- | --------- | ---------------------------------- |\n| Text synthesis         | Yes       |                                    |\n| AI response synthesis  | Yes       | Set useAiResponse: true          |\n| HD quality             | Yes       | quality: \"hd\" maps to tts-1-hd |\n| Speed control          | Yes       | 0.25 – 4.0                         |\n| Voice selection        | Yes       | 6 neural voices   ","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI TTS Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"3124582a399e5bb98ce8924efecb3558d0161c1f16faae7fdedbfc2e182e34a0","title":"OpenAI TTS Provider Guide","url":"/docs/getting-started/providers/openai-tts#openai-tts-provider-guide","content":"High-quality neural text-to-speech with six distinct voices and HD quality option","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI TTS Provider Guide","lvl2":"OpenAI TTS Provider Guide","lvl3":""}},
 {"objectID":"b0ec9b5b8b5214c2b809622c60e20e66d39e7e3db9353215ab1f5de78036fac8","title":"Overview","url":"/docs/getting-started/providers/openai-tts#overview","content":"NeuroLink integrates OpenAI's Text-to-Speech API, giving you access to six expressive neural voices across two model tiers. The standard model (tts-1) optimises for low latency, while the HD model (tts-1-hd) delivers higher audio fidelity for production use cases such as podcasts, voice assistants, and narration.\n\nOpenAI TTS works with any NeuroLink text generation call — you can synthesise the raw prompt directly or synthesise the AI-generated response, controlled by the useAiResponse flag.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI TTS Provider Guide","lvl2":"Overview","lvl3":""}},
@@ -6857,7 +6862,7 @@
 {"objectID":"e2d66c098b27a18e31a0985c2b41817ec2896b37532ec07b38e09f1ed760c1b7","title":"\"OpenAI TTS request timed out after 30 seconds\"","url":"/docs/getting-started/providers/openai-tts#openai-tts-request-timed-out-after-30-seconds","content":"A network issue or overloaded API caused the request to time out. Retry the request — the error is marked retriable by NeuroLink's error system.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI TTS Provider Guide","lvl2":"\"OpenAI TTS request timed out after 30 seconds\"","lvl3":""}},
 {"objectID":"9a54f6363be2937d7c7f9328a1e90b4e1945f5d0c31f3e66477c6468d131b099","title":"Audio sounds distorted at high speed","url":"/docs/getting-started/providers/openai-tts#audio-sounds-distorted-at-high-speed","content":"Speeds above 2.0 can introduce artifacts. Use speed: 1.0 – 1.5 for natural-sounding output.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI TTS Provider Guide","lvl2":"Audio sounds distorted at high speed","lvl3":""}},
 {"objectID":"f98d39032204ab3aed29744d24a667388d714a7e1214577e166e2d377469d80e","title":"See Also","url":"/docs/getting-started/providers/openai-tts#see-also","content":"TTS Integration Guide — complete multi-provider TTS reference\nAudio Input (STT) — speech-to-text counterpart\nOpenAI Provider Guide — full OpenAI text generation provider\nElevenLabs Provider Guide — alternative TTS provider with voice cloning\n\nNeed Help? Join the GitHub Discussions or open an issue.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI TTS Provider Guide","lvl2":"See Also","lvl3":""}},
-{"objectID":"988dc021ae1e24ac074ba6ccd12141310a44053f1327466076993828c5bf0d61","title":"OpenAI Provider Guide","url":"/docs/getting-started/providers/openai","content":"OpenAI Provider Guide\n\nAccess GPT-5.4, GPT-5, GPT-4o, o-series reasoning models, and embedding models through the OpenAI API\n\nOverview\n\nOpenAI provides API access to the GPT model family, including the GPT-6 family, the GPT-5.4 series, GPT-5 series, GPT-4o multimodal models, and o-series reasoning models. NeuroLink talks to the OpenAI HTTP API directly — generation, streaming, tool calling, vision and embeddings are all served by NeuroLink's own client, with no third-party model SDK in the path.\n\nKey Benefits\nGPT-5.4 Series: Flagship models (March 2026) with 1.05M context (gpt-5.4, gpt-5.4-pro; mini and nano 400K)\nGPT-5 Series: Flagship models with up to 400K context windows\nGPT-4.1 Series: 1M context window models for large document processing\nGPT-4o: Multimodal model with vision support\no-Series Reasoning: o3, o3-pro, and o4-mini for deep reasoning tasks\nEmbeddings: text-embedding-3-small and other embedding models\nTool/Function Calling: Full support for agent workflows\nStreaming: Real-time streaming responses with tool execution\nProxy Support: Route requests through HTTP/HTTPS/SOCKS proxies\n\nProvider Aliases\n\nYou can reference this provider using any of the following names:\n\n| Alias     | Usage                       |\n| --------- | --------------------------- |\n| openai  | Canonical provider name     |\n| gpt     | Short alias for convenience |\n| chatgpt | Alternative alias           |\n\nThese aliases are registered in src/lib/factories/providerRegistry.ts.\n\nQuick Start\nGet Your API Key\nVisit platform.openai.com/api-keys\nSign in or create an account\nClick Create new secret key\nCopy your new API key (starts with sk-)\nConfigure Environment\n\nAdd to your .env file:\n\nOPENAI_API_KEY also configures decisions. The same variable configures the\nOpenAI Decisions provider (openai-decisions, which\nserves decide()). So setting it for chat also lets NeuroLink's built-in\nfeatures send decision requests to OpenAI when no TypeSafe, Laya, XOR,\nPerplexity or Cloudflare Clef provider is configured, and context compaction\nneeds no other opt-in to do so. Those are paid calls, and they fail open. To\navoid that, pass the key through the SDK as credentials.openai.apiKey instead\nof the environment, and keep it out of .env too: the SDK and the CLI load that\nfile into the environment, so an OPENAI_API_KEY line like the one above would\nstill configure decisions. See\nWhat is sent to OpenAI and\nOne key, two providers.\nIf you point the chat provider at a gateway (LiteLLM, vLLM and the like) with\nOPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL, the key is the gateway's and is not\nused for decisions by default: while any of them names a host other than\nhttps://api.openai.com, NeuroLink's built-in features send nothing to OpenAI's\nDecisions API (only a call that names openai-decisions does, with that key). See\nA gateway's key.\nTest the Setup\n\nSupported Models\n\nAvailable Models (from OpenAIModels enum)\n\n| Enum Key              | Model ID              | Series       | Context Window | Notes                    |\n| --------------------- | --------------------- | ------------ | -------------- | ------------------------ |\n| GPT_6_ASTRA         | gpt-6-astra         | GPT-6        | 1.05M          | New (September 2026) |\n| GPT_6_SOL           | gpt-6-sol           | GPT-6        | 1.05M          | New (September 2026) |\n| GPT_6_LUNA          | gpt-6-luna          | GPT-6        | 1.05M          | New (September 2026) |\n| GPT_5_4             | gpt-5.4             | GPT-5.4      | 1.05M          | New (March 2026)     |\n| GPT_5_4_MINI        | gpt-5.4-mini        | GPT-5.4      | 400K           | New (March 2026)     |\n| GPT_5_4_NANO        | gpt-5.4-nano        | GPT-5.4      | 400K           | New (March 2026)     |\n| GPT_5_3_CODEX       | gpt-5.3-codex       | GPT-5.3      | 400K           |                          |\n| GPT_5_2             | gpt-5.2             | GPT-5.2      | 400K           |                          |\n| GPT_5_2_CHAT_LATEST | gpt-5.2-chat-latest | GPT-5.2      | 128K           |                          |\n| GPT_5_2_PRO         | gpt-5.2-pro         | GPT-5.2      | 400K           |                          |\n| GPT_5_2_CODEX       | gpt-5.2-codex       | GPT-5.2      | 400K           |                          |\n| GPT_5_1             | gpt-5.1             | GPT-5.1      | 400K           |                          |\n| GPT_5_1_CHAT_LATEST | gpt-5.1-chat-latest | GPT-5.1      | 128K           |                          |\n| GPT_5_1_CODEX       | gpt-5.1-codex       | GPT-5.1      | 400K           |                          |\n| GPT_5_1_CODEX_MAX   | gpt-5.1-codex-max   | GPT-5.1      | 400K           |                          |\n| GPT_5_1_CODEX_MINI  | gpt-5.1-codex-mini  | GPT-5.1      | 400K           |                          |\n| GPT_5               | gpt-5               | GPT-5        | 400K           |                          |\n| GPT_5_MINI          | gpt-5-mini          | GPT-5        | 400K           |       ","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Provider Guide","lvl2":"","lvl3":""}},
+{"objectID":"988dc021ae1e24ac074ba6ccd12141310a44053f1327466076993828c5bf0d61","title":"OpenAI Provider Guide","url":"/docs/getting-started/providers/openai","content":"OpenAI Provider Guide\n\nAccess GPT-5.4, GPT-5, GPT-4o, o-series reasoning models, and embedding models through the OpenAI API\n\nOverview\n\nOpenAI provides API access to the GPT model family, including the GPT-6 family, the GPT-5.4 series, GPT-5 series, GPT-4o multimodal models, and o-series reasoning models. NeuroLink talks to the OpenAI HTTP API directly — generation, streaming, tool calling, vision and embeddings are all served by NeuroLink's own client, with no third-party model SDK in the path.\n\nKey Benefits\nGPT-5.4 Series: Flagship models (March 2026) with 1.05M context (gpt-5.4, gpt-5.4-pro; mini and nano 400K)\nGPT-5 Series: Flagship models with up to 400K context windows\nGPT-4.1 Series: 1M context window models for large document processing\nGPT-4o: Multimodal model with vision support\no-Series Reasoning: o3, o3-pro, and o4-mini for deep reasoning tasks\nEmbeddings: text-embedding-3-small and other embedding models\nTool/Function Calling: Full support for agent workflows\nStreaming: Real-time streaming responses with tool execution\nProxy Support: Route requests through HTTP/HTTPS/SOCKS proxies\n\nProvider Aliases\n\nYou can reference this provider using any of the following names:\n\n| Alias     | Usage                       |\n| --------- | --------------------------- |\n| openai  | Canonical provider name     |\n| gpt     | Short alias for convenience |\n| chatgpt | Alternative alias           |\n\nThese aliases are registered in src/lib/factories/providerRegistry.ts.\n\nQuick Start\nGet Your API Key\nVisit platform.openai.com/api-keys\nSign in or create an account\nClick Create new secret key\nCopy your new API key (starts with sk-)\nConfigure Environment\n\nAdd to your .env file:\n\nOPENAI_API_KEY also configures decisions. The same variable configures the\nOpenAI Decisions provider (openai-decisions, which\nserves decide()). So setting it for chat also lets NeuroLink's built-in\nfeatures send decision requests to OpenAI when no TypeSafe, Laya, XOR, SubQ,\nPerplexity or Cloudflare Clef provider is configured, and context compaction\nneeds no other opt-in to do so. Those are paid calls, and they fail open. To\navoid that, pass the key through the SDK as credentials.openai.apiKey instead\nof the environment, and keep it out of .env too: the SDK and the CLI load that\nfile into the environment, so an OPENAI_API_KEY line like the one above would\nstill configure decisions. See\nWhat is sent to OpenAI and\nOne key, two providers.\nIf you point the chat provider at a gateway (LiteLLM, vLLM and the like) with\nOPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL, the key is the gateway's and is not\nused for decisions by default: while any of them names a host other than\nhttps://api.openai.com, NeuroLink's built-in features send nothing to OpenAI's\nDecisions API (only a call that names openai-decisions does, with that key). See\nA gateway's key.\nTest the Setup\n\nSupported Models\n\nAvailable Models (from OpenAIModels enum)\n\n| Enum Key              | Model ID              | Series       | Context Window | Notes                    |\n| --------------------- | --------------------- | ------------ | -------------- | ------------------------ |\n| GPT_6_ASTRA         | gpt-6-astra         | GPT-6        | 1.05M          | New (September 2026) |\n| GPT_6_SOL           | gpt-6-sol           | GPT-6        | 1.05M          | New (September 2026) |\n| GPT_6_LUNA          | gpt-6-luna          | GPT-6        | 1.05M          | New (September 2026) |\n| GPT_5_4             | gpt-5.4             | GPT-5.4      | 1.05M          | New (March 2026)     |\n| GPT_5_4_MINI        | gpt-5.4-mini        | GPT-5.4      | 400K           | New (March 2026)     |\n| GPT_5_4_NANO        | gpt-5.4-nano        | GPT-5.4      | 400K           | New (March 2026)     |\n| GPT_5_3_CODEX       | gpt-5.3-codex       | GPT-5.3      | 400K           |                          |\n| GPT_5_2             | gpt-5.2             | GPT-5.2      | 400K           |                          |\n| GPT_5_2_CHAT_LATEST | gpt-5.2-chat-latest | GPT-5.2      | 128K           |                          |\n| GPT_5_2_PRO         | gpt-5.2-pro         | GPT-5.2      | 400K           |                          |\n| GPT_5_2_CODEX       | gpt-5.2-codex       | GPT-5.2      | 400K           |                          |\n| GPT_5_1             | gpt-5.1             | GPT-5.1      | 400K           |                          |\n| GPT_5_1_CHAT_LATEST | gpt-5.1-chat-latest | GPT-5.1      | 128K           |                          |\n| GPT_5_1_CODEX       | gpt-5.1-codex       | GPT-5.1      | 400K           |                          |\n| GPT_5_1_CODEX_MAX   | gpt-5.1-codex-max   | GPT-5.1      | 400K           |                          |\n| GPT_5_1_CODEX_MINI  | gpt-5.1-codex-mini  | GPT-5.1      | 400K           |                          |\n| GPT_5               | gpt-5               | GPT-5        | 400K           |                          |\n| GPT_5_MINI          | gpt-5-mini          | GPT-5        | 400K           | ","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"09a56d9dc91320ff37eda19c92d08529dc6471dd3b273124ca539bf8c68cf8eb","title":"OpenAI Provider Guide","url":"/docs/getting-started/providers/openai#openai-provider-guide","content":"Access GPT-5.4, GPT-5, GPT-4o, o-series reasoning models, and embedding models through the OpenAI API","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Provider Guide","lvl2":"OpenAI Provider Guide","lvl3":""}},
 {"objectID":"00a9eec16078fe5bf845c9876ae67265e060b64f3150a402681173526d5f9037","title":"Overview","url":"/docs/getting-started/providers/openai#overview","content":"OpenAI provides API access to the GPT model family, including the GPT-6 family, the GPT-5.4 series, GPT-5 series, GPT-4o multimodal models, and o-series reasoning models. NeuroLink talks to the OpenAI HTTP API directly — generation, streaming, tool calling, vision and embeddings are all served by NeuroLink's own client, with no third-party model SDK in the path.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Provider Guide","lvl2":"Overview","lvl3":""}},
 {"objectID":"7ef3ddd9cd1195ba96e4a45b2019a85d9ef0f4aa77585f4cff3f58140d40ea8e","title":"Key Benefits","url":"/docs/getting-started/providers/openai#key-benefits","content":"GPT-5.4 Series: Flagship models (March 2026) with 1.05M context (gpt-5.4, gpt-5.4-pro; mini and nano 400K)\nGPT-5 Series: Flagship models with up to 400K context windows\nGPT-4.1 Series: 1M context window models for large document processing\nGPT-4o: Multimodal model with vision support\no-Series Reasoning: o3, o3-pro, and o4-mini for deep reasoning tasks\nEmbeddings: text-embedding-3-small and other embedding models\nTool/Function Calling: Full support for agent workflows\nStreaming: Real-time streaming responses with tool execution\nProxy Support: Route requests through HTTP/HTTPS/SOCKS proxies","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Provider Guide","lvl2":"Key Benefits","lvl3":""}},
@@ -6865,7 +6870,7 @@
 {"objectID":"769a8aea5668e683c64e455fc484af87c547264189eaed08c2ff047dfc535167","title":"1. Get Your API Key","url":"/docs/getting-started/providers/openai#1-get-your-api-key","content":"Visit platform.openai.com/api-keys\nSign in or create an account\nClick Create new secret key\nCopy your new API key (starts with sk-)","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Provider Guide","lvl2":"1. Get Your API Key","lvl3":""}},
 {"objectID":"c2df7c52c3715cfca6e8c91fd43d047a5c50e2fd573e9ca5b78784aa4b98068c","title":"2. Configure Environment","url":"/docs/getting-started/providers/openai#2-configure-environment","content":"Add to your .env file:\n\n`bash","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Provider Guide","lvl2":"2. Configure Environment","lvl3":""}},
 {"objectID":"877d2901da12fab35c77fd0fca652f1e43e6ddea39645924dac938ef99b4463d","title":"Required: Your OpenAI API key","url":"/docs/getting-started/providers/openai#required-your-openai-api-key","content":"OPENAIAPIKEY=sk-your-key-here","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Provider Guide","lvl2":"Required: Your OpenAI API key","lvl3":""}},
-{"objectID":"e1ea9f50a3405fe5b61c5bf8655daf66c0f353f33b9dff00f4181be2a1258e53","title":"Optional: Override default model (defaults to gpt-4o-mini)","url":"/docs/getting-started/providers/openai#optional-override-default-model-defaults-to-gpt-4o-mini","content":"OPENAI_MODEL=gpt-4o\n\n\n> **OPENAIAPIKEY also configures decisions.** The same variable configures the\n> [OpenAI Decisions provider](/docs/getting-started/providers/openai-decisions) (openai-decisions, which\n> serves decide()). So setting it for chat also lets NeuroLink's built-in\n> features send decision requests to OpenAI when no TypeSafe, Laya, XOR,\n> Perplexity or Cloudflare Clef provider is configured, and context compaction\n> needs no other opt-in to do so. Those are paid calls, and they fail open. To\n> avoid that, pass the key through the SDK as credentials.openai.apiKey instead\n> of the environment, and keep it out of .env too: the SDK and the CLI load that\n> file into the environment, so an OPENAIAPIKEY line like the one above would\n> still configure decisions. See\n> [What is sent to OpenAI](/docs/getting-started/providers/openai-decisions#what-is-sent-to-openai) and\n> [One key, two providers](/docs/getting-started/providers/openai-decisions#one-key-two-providers).\n>\n> If you point the chat provider at a gateway (LiteLLM, vLLM and the like) with\n> OPENAIBASEURL, OPENAIAPIBASE or credentials.openai.baseURL, the key is the gateway's and is **not**\n> used for decisions by default: while any of them names a host other than\n> https://api.openai.com, NeuroLink's built-in features send nothing to OpenAI's\n> Decisions API (only a call that names openai-decisions` does, with that key). See\nA gateway's key.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Provider Guide","lvl2":"Optional: Override default model (defaults to gpt-4o-mini)","lvl3":""}},
+{"objectID":"e1ea9f50a3405fe5b61c5bf8655daf66c0f353f33b9dff00f4181be2a1258e53","title":"Optional: Override default model (defaults to gpt-4o-mini)","url":"/docs/getting-started/providers/openai#optional-override-default-model-defaults-to-gpt-4o-mini","content":"OPENAI_MODEL=gpt-4o\n\n\n> **OPENAIAPIKEY also configures decisions.** The same variable configures the\n> [OpenAI Decisions provider](/docs/getting-started/providers/openai-decisions) (openai-decisions, which\n> serves decide()). So setting it for chat also lets NeuroLink's built-in\n> features send decision requests to OpenAI when no TypeSafe, Laya, XOR, SubQ,\n> Perplexity or Cloudflare Clef provider is configured, and context compaction\n> needs no other opt-in to do so. Those are paid calls, and they fail open. To\n> avoid that, pass the key through the SDK as credentials.openai.apiKey instead\n> of the environment, and keep it out of .env too: the SDK and the CLI load that\n> file into the environment, so an OPENAIAPIKEY line like the one above would\n> still configure decisions. See\n> [What is sent to OpenAI](/docs/getting-started/providers/openai-decisions#what-is-sent-to-openai) and\n> [One key, two providers](/docs/getting-started/providers/openai-decisions#one-key-two-providers).\n>\n> If you point the chat provider at a gateway (LiteLLM, vLLM and the like) with\n> OPENAIBASEURL, OPENAIAPIBASE or credentials.openai.baseURL, the key is the gateway's and is **not**\n> used for decisions by default: while any of them names a host other than\n> https://api.openai.com, NeuroLink's built-in features send nothing to OpenAI's\n> Decisions API (only a call that names openai-decisions` does, with that key). See\nA gateway's key.","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Provider Guide","lvl2":"Optional: Override default model (defaults to gpt-4o-mini)","lvl3":""}},
 {"objectID":"5e7e8211d2fb3a2017c349fcc7d112734942bf738f8202fcf05163a1fa3b9249","title":"3. Test the Setup","url":"/docs/getting-started/providers/openai#3-test-the-setup","content":"`bash","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Provider Guide","lvl2":"3. Test the Setup","lvl3":""}},
 {"objectID":"51393d2a6f860f549650441d19ba86353e5feb0d267b47894b1cd40e16d7e3d0","title":"Quick generation","url":"/docs/getting-started/providers/openai#quick-generation","content":"pnpm run cli -- generate \"Hello from GPT!\" \\\n  --provider openai","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Provider Guide","lvl2":"Quick generation","lvl3":""}},
 {"objectID":"6102cd238f922f95762068dd3d2b3950e839a87ac388ee96bb805db79c47d00e","title":"Use specific model","url":"/docs/getting-started/providers/openai#use-specific-model","content":"pnpm run cli -- generate \"Write a haiku about AI\" \\\n  --provider openai \\\n  --model \"gpt-4o\"","hierarchy":{"lvl0":"Getting Started","lvl1":"OpenAI Provider Guide","lvl2":"Use specific model","lvl3":""}},
@@ -6980,16 +6985,16 @@
 {"objectID":"106d04b8baec018f872ce608c3b0dd33f8b3d43ada85072831a2e9ab9c64c46b","title":"Verification status","url":"/docs/getting-started/providers/pareto-inference#verification-status","content":"Tier-2 onboarding requires evidence before a provider is accepted, and\npnpm run verify:provider-onboarding gates it in CI. This is what the\ncatalog records for Pareto Inference — and, just as importantly, what it\ndoes not yet record:\n\n| Probe                 | Result                                                                                                                                                                                                                                                                                                                                                                                             |\n| --------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Roster                | unauthenticated GET /v1/models, HTTP 200, 1 model (z-ai/glm-5.3-flash), 2026-09-28. No API key was used or required for this call.                                                                                                                                                                                                                                                         |\n| Auth rejection        | Not probed. Would require a POST with an invalid key, which this credential-free onboarding pass does not send. Pareto's docs describe 401 as \"the chat API key is missing, invalid, or no longer active,\" with no example error body.                                                                                                                                                           |\n| Live capability sweep | Not run. evidence.liveMatrix is null.","hierarchy":{"lvl0":"Getting Started","lvl1":"Pareto Inference Provider Guide","lvl2":"Verification status","lvl3":""}},
 {"objectID":"b844edaf09c93fb3089bbf0a71254c5039dedf111e2b59563321484496862aa4","title":"Troubleshooting","url":"/docs/getting-started/providers/pareto-inference#troubleshooting","content":"| Symptom                                 | Cause                                                                | Fix                                                                                                                                             |\n| --------------------------------------- | -------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- |\n| Invalid Pareto Inference API key      | PARETO_INFERENCE_API_KEY unset, wrong, or expired                  | Pareto documents 401 as \"the chat API key is missing, invalid, or no longer active\" — check or replace the key in the dashboard                 |\n| 429 with credit_exhausted             | Prepaid balance is empty                                             | Buy more prepaid credits                                                                                                                        |\n| 429 with credit_insufficient          | Held credits (reserved for max_tokens) are below what's needed     | Lower max_tokens, or buy more credits                                                                                                         |\n| 429 with model_capacity               | The model itself is at capacity — not an account-level limit         | Honor the Retry-After header and retry                                                                                                        |\n| 502 mid-stream                          | The model request failed after the stream started                    | Retry the request; check whether a tool call already ran before retrying                                                                        |\n| 503                                     | Temporary Pareto-side failure (e.g. maintenance, GPU outage)         | Wait and retry; Retry-After gives the delay in seconds","hierarchy":{"lvl0":"Getting Started","lvl1":"Pareto Inference Provider Guide","lvl2":"Troubleshooting","lvl3":""}},
 {"objectID":"57b8457eff592f229d3c0f0fb7c73148341ecffd72bb29c15d73f1471b1fd886","title":"See also","url":"/docs/getting-started/providers/pareto-inference#see-also","content":"Provider setup overview\nAll providers\nTier-2 onboarding — how this provider's JSON becomes a working integration\nProvider feature compatibility","hierarchy":{"lvl0":"Getting Started","lvl1":"Pareto Inference Provider Guide","lvl2":"See also","lvl3":""}},
-{"objectID":"daf00919862e4bbed4b3fc1082bea17843a987e5e3970cefed2c95537b22cceb","title":"Perplexity Decisions Provider Guide","url":"/docs/getting-started/providers/perplexity-decider","content":"Perplexity Decisions Provider Guide\n\nA provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR,\nCloudflare Clef and OpenAI Decisions,\nfrom Perplexity's hosted Decisions API, and it also reads images. It emits no\ntext at all.\n\nThis is not the Perplexity text provider (perplexity, the\nSonar models), which serves generate() and stream(). The two share one API\nkey, which has a consequence worth reading before you set it. See\nWhat is sent to Perplexity and\nOne key, two providers.\n\nOverview\n\npplx-decider-v1-27b is Perplexity's decision model. You send one state\nplus named, typed questions, and optionally images. The model answers every\nquestion in a single batched pass and returns a typed answer for each, with a\nprobability instead of a sentence. NeuroLink calls Perplexity's public endpoint,\nso a key alone configures it: there is no base URL to set.\n\nKey Facts\n\n|                          |                                                                                                                   |\n| ------------------------ | ----------------------------------------------------------------------------------------------------------------- |\n| Provider id              | perplexity-decider (no aliases) — distinct from perplexity, the Sonar text provider                           |\n| Inference type           | decide only                                                                                                     |\n| Model                    | pplx-decider-v1-27b; the API answers 400 to a missing or unknown model. PERPLEXITY_DECIDER_MODEL sets it      |\n| Key                      | PERPLEXITY_API_KEY, shared with the Perplexity text provider, or credentials.perplexityDecider.apiKey         |\n| Endpoint                 | https://api.perplexity.ai/v1/decisions; PERPLEXITY_DECIDER_BASE_URL can name another origin                   |\n| Media                    | images (PNG, JPEG or WebP), up to 8 per request; no video                                                         |\n| Questions per request    | up to 128                                                                                                         |\n| Server input ceiling     | under 262,144 tokens (state, questions and images); more is refused with an explicit 400, never cut off silently  |\n| Images and that ceiling  | billed as input tokens, and counted toward the ceiling at one token per 32 × 32 tile (measured)                   |\n| NeuroLink's state window | 100,000 estimated tokens: a deliberate local limit, not the server's                                              |\n| Cost                     | $0.04 per million input tokens (image tokens included); output tokens are free. Perplexity's documented price     |\n| Default timeout          | 10 seconds plus 100 ms for each question (10.1 s for one, 22.8 s for 128); timeoutMs or --timeout replaces it |\n| Precedence               | used automatically when TypeSafe, Laya, XOR and SubQ are not configured                                           |\n\nEach limit, token rate and latency in this guide is either measured on a real\naccount in October 2026 or taken from Perplexity's documentation, and says\nwhich where it appears. Measured: the 128-question and 8-image caps, the\n262,144-token input ceiling for the state, the questions and the images (nothing\nis cut off silently), characters per token for each kind of text, latency by\ninput size and by question count, what an image costs, which image sizes stall\n(the 2,048-tile rule, checked at its edge: see Images), and how the\nAPI answers a burst of requests. Documented and not tested: the 32 MiB request\nbody, the price, and the option and level counts. Perplexity documents a limit of\n10 requests per second for the account's tier (every organization, on every\nplan, with a token limit on large bursts), and a burst test on the one account\ntested saw the request limit act.\n\nQuick Start\nGet a key\n\nCreate one in the Perplexity console. Any\nPerplexity API key works.\nConfigure\n\nSet the key in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:\n\ncredentials.perplexityDecider is its own slice. A key passed as\ncredentials.perplexity belongs to the text provider and does not configure\nthis one. PERPLEXITY_BASE_URL, which moves the text provider, is not read here;\nthe override for this provider is PERPLEXITY_DECIDER_BASE_URL, and a trailing\n/v1 on it is accepted. It must be https://; http:// is accepted only for\nlocalhost, 127.0.0.1 and [::1], because every request carries the key.\nUse it\n\nNeuroLink's boolean question is sent to the API as its noul type, and the\nanswer is read back as a boolean. A boolean answer carries a probability and\nno confidence of its own; a choice or score answer carries the confidence the\nAPI reports, which Perplexity describes as the model's own certainty estimate,\nnot the","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"","lvl3":""}},
-{"objectID":"cd579839c9ecb7a84dee24ccd591e76f90eb1f05b384a35488c7603f72178d38","title":"Perplexity Decisions Provider Guide","url":"/docs/getting-started/providers/perplexity-decider#perplexity-decisions-provider-guide","content":"A provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR,\nCloudflare Clef and OpenAI Decisions,\nfrom Perplexity's hosted Decisions API, and it also reads images. It emits no\ntext at all.\n\nThis is not the Perplexity text provider (perplexity, the\nSonar models), which serves generate() and stream(). The two share one API\nkey, which has a consequence worth reading before you set it. See\nWhat is sent to Perplexity and\nOne key, two providers.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"Perplexity Decisions Provider Guide","lvl3":""}},
+{"objectID":"daf00919862e4bbed4b3fc1082bea17843a987e5e3970cefed2c95537b22cceb","title":"Perplexity Decisions Provider Guide","url":"/docs/getting-started/providers/perplexity-decider","content":"Perplexity Decisions Provider Guide\n\nA provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR, SubQ,\nCloudflare Clef, OpenAI Decisions\nand ScaleDown, from Perplexity's hosted Decisions API, and it also reads images. It emits no\ntext at all.\n\nThis is not the Perplexity text provider (perplexity, the\nSonar models), which serves generate() and stream(). The two share one API\nkey, which has a consequence worth reading before you set it. See\nWhat is sent to Perplexity and\nOne key, two providers.\n\nOverview\n\npplx-decider-v1-27b is Perplexity's decision model. You send one state\nplus named, typed questions, and optionally images. The model answers every\nquestion in a single batched pass and returns a typed answer for each, with a\nprobability instead of a sentence. NeuroLink calls Perplexity's public endpoint,\nso a key alone configures it: there is no base URL to set.\n\nKey Facts\n\n|                          |                                                                                                                   |\n| ------------------------ | ----------------------------------------------------------------------------------------------------------------- |\n| Provider id              | perplexity-decider (no aliases) — distinct from perplexity, the Sonar text provider                           |\n| Inference type           | decide only                                                                                                     |\n| Model                    | pplx-decider-v1-27b; the API answers 400 to a missing or unknown model. PERPLEXITY_DECIDER_MODEL sets it      |\n| Key                      | PERPLEXITY_API_KEY, shared with the Perplexity text provider, or credentials.perplexityDecider.apiKey         |\n| Endpoint                 | https://api.perplexity.ai/v1/decisions; PERPLEXITY_DECIDER_BASE_URL can name another origin                   |\n| Media                    | images (PNG, JPEG or WebP), up to 8 per request; no video                                                         |\n| Questions per request    | up to 128                                                                                                         |\n| Server input ceiling     | under 262,144 tokens (state, questions and images); more is refused with an explicit 400, never cut off silently  |\n| Images and that ceiling  | billed as input tokens, and counted toward the ceiling at one token per 32 × 32 tile (measured)                   |\n| NeuroLink's state window | 100,000 estimated tokens: a deliberate local limit, not the server's                                              |\n| Cost                     | $0.04 per million input tokens (image tokens included); output tokens are free. Perplexity's documented price     |\n| Default timeout          | 10 seconds plus 100 ms for each question (10.1 s for one, 22.8 s for 128); timeoutMs or --timeout replaces it |\n| Precedence               | used automatically when TypeSafe, Laya, XOR and SubQ are not configured                                           |\n\nEach limit, token rate and latency in this guide is either measured on a real\naccount in October 2026 or taken from Perplexity's documentation, and says\nwhich where it appears. Measured: the 128-question and 8-image caps, the\n262,144-token input ceiling for the state, the questions and the images (nothing\nis cut off silently), characters per token for each kind of text, latency by\ninput size and by question count, what an image costs, which image sizes stall\n(the 2,048-tile rule, checked at its edge: see Images), and how the\nAPI answers a burst of requests. Documented and not tested: the 32 MiB request\nbody, the price, and the option and level counts. Perplexity documents a limit of\n10 requests per second for the account's tier (every organization, on every\nplan, with a token limit on large bursts), and a burst test on the one account\ntested saw the request limit act.\n\nQuick Start\nGet a key\n\nCreate one in the Perplexity console. Any\nPerplexity API key works.\nConfigure\n\nSet the key in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:\n\ncredentials.perplexityDecider is its own slice. A key passed as\ncredentials.perplexity belongs to the text provider and does not configure\nthis one. PERPLEXITY_BASE_URL, which moves the text provider, is not read here;\nthe override for this provider is PERPLEXITY_DECIDER_BASE_URL, and a trailing\n/v1 on it is accepted. It must be https://; http:// is accepted only for\nlocalhost, 127.0.0.1 and [::1], because every request carries the key.\nUse it\n\nNeuroLink's boolean question is sent to the API as its noul type, and the\nanswer is read back as a boolean. A boolean answer carries a probability and\nno confidence of its own; a choice or score answer carries the confidence the\nAPI reports, which Perplexity describes as the model's own certainty ","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"","lvl3":""}},
+{"objectID":"cd579839c9ecb7a84dee24ccd591e76f90eb1f05b384a35488c7603f72178d38","title":"Perplexity Decisions Provider Guide","url":"/docs/getting-started/providers/perplexity-decider#perplexity-decisions-provider-guide","content":"A provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR, SubQ,\nCloudflare Clef, OpenAI Decisions\nand ScaleDown, from Perplexity's hosted Decisions API, and it also reads images. It emits no\ntext at all.\n\nThis is not the Perplexity text provider (perplexity, the\nSonar models), which serves generate() and stream(). The two share one API\nkey, which has a consequence worth reading before you set it. See\nWhat is sent to Perplexity and\nOne key, two providers.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"Perplexity Decisions Provider Guide","lvl3":""}},
 {"objectID":"0d2eec9390e1cc5078e5419138c190cc8dc524c5ebd8972a9c7afa9625dae43a","title":"Overview","url":"/docs/getting-started/providers/perplexity-decider#overview","content":"pplx-decider-v1-27b is Perplexity's decision model. You send one state\nplus named, typed questions, and optionally images. The model answers every\nquestion in a single batched pass and returns a typed answer for each, with a\nprobability instead of a sentence. NeuroLink calls Perplexity's public endpoint,\nso a key alone configures it: there is no base URL to set.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"Overview","lvl3":""}},
 {"objectID":"c9690ba724ce4205d66ac4d75981884b735170b152767047afcc0d406d5e7c8d","title":"Key Facts","url":"/docs/getting-started/providers/perplexity-decider#key-facts","content":"|                          |                                                                                                                   |\n| ------------------------ | ----------------------------------------------------------------------------------------------------------------- |\n| Provider id              | perplexity-decider (no aliases) — distinct from perplexity, the Sonar text provider                           |\n| Inference type           | decide only                                                                                                     |\n| Model                    | pplx-decider-v1-27b; the API answers 400 to a missing or unknown model. PERPLEXITY_DECIDER_MODEL sets it      |\n| Key                      | PERPLEXITY_API_KEY, shared with the Perplexity text provider, or credentials.perplexityDecider.apiKey         |\n| Endpoint                 | https://api.perplexity.ai/v1/decisions; PERPLEXITY_DECIDER_BASE_URL can name another origin                   |\n| Media                    | images (PNG, JPEG or WebP), up to 8 per request; no video                                                         |\n| Questions per request    | up to 128                                                                                                         |\n| Server input ceiling     | under 262,144 tokens (state, questions and images); more is refused with an explicit 400, never cut off silently  |\n| Images and that ceiling  | billed as input tokens, and counted toward the ceiling at one token per 32 × 32 tile (measured)                   |\n| NeuroLink's state window | 100,000 estimated tokens: a deliberate local limit, not the server's                                              |\n| Cost                     | $0.04 per million input tokens (image tokens included); output tokens are free.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"Key Facts","lvl3":""}},
 {"objectID":"467fb7efd250abc71afba6a9535ee164e671526dc06ff59407a827bda872b268","title":"1. Get a key","url":"/docs/getting-started/providers/perplexity-decider#1-get-a-key","content":"Create one in the Perplexity console. Any\nPerplexity API key works.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"1. Get a key","lvl3":""}},
 {"objectID":"bd7509f708e016023be201690ca4cc963bc77551d7152b34c636ae628ea5e497","title":"2. Configure","url":"/docs/getting-started/providers/perplexity-decider#2-configure","content":"Set the key in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:\n\ncredentials.perplexityDecider is its own slice. A key passed as\ncredentials.perplexity belongs to the text provider and does not configure\nthis one. PERPLEXITY_BASE_URL, which moves the text provider, is not read here;\nthe override for this provider is PERPLEXITY_DECIDER_BASE_URL, and a trailing\n/v1 on it is accepted. It must be https://; http:// is accepted only for\nlocalhost, 127.0.0.1 and [::1], because every request carries the key.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"2. Configure","lvl3":""}},
 {"objectID":"c3dbe3c712926dfb07b0d907370532df8bc3c98a5e81f4b46784a037ac0f1e0b","title":"3. Use it","url":"/docs/getting-started/providers/perplexity-decider#3-use-it","content":"NeuroLink's boolean question is sent to the API as its noul type, and the\nanswer is read back as a boolean. A boolean answer carries a probability and\nno confidence of its own; a choice or score answer carries the confidence the\nAPI reports, which Perplexity describes as the model's own certainty estimate,\nnot the top probability. Perplexity's API reference and quickstart do not call\nthat confidence calibrated, and NeuroLink did not measure whether it is, so tune\na threshold on it against your own labelled data. (The Hugging Face model card\nfor the open-weights model describes its output as calibrated probabilities.)\nPerplexity also notes that identical requests occasionally differ in the second\ndecimal place. Repeating a request returned the same numbers in NeuroLink's\nprobe, but leave some margin when you set a threshold anyway.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"3. Use it","lvl3":""}},
 {"objectID":"6f3f0a67d6c85ce65d1d7d95a7016987ad0423b545c9f849bcde853edd0dd20f","title":"From the CLI","url":"/docs/getting-started/providers/perplexity-decider#from-the-cli","content":"The CLI reads the same environment variables:\n\nThe text output ends with a Cost: line, priced from the input tokens the API\nreports.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"From the CLI","lvl3":""}},
-{"objectID":"e1b9161659de68a10277e69ce53077f2aed241892d783054f01471fb793024bd","title":"Images","url":"/docs/getting-started/providers/perplexity-decider#images","content":"Perplexity reads images alongside the state. TypeSafe and Laya do not read\nmedia. XOR reads images and a video; Perplexity reads images only, and so do\nCloudflare Clef, up to 4 of them, and\nOpenAI Decisions. In the probe of\nPerplexity's model it did read an image: asked red versus blue about one, it\nanswered correctly at about 0.98 confidence.\n\nFrom the CLI, pass --image <path>, repeated for several:\n\nOn the wire the images go inside state: NeuroLink sends state as an array,\nyour state first and then one OpenAI-style image_url part per image, each a\nbase64 data URL. The API has no separate images field and rejects unknown\nfields. An image can also be the whole input (an image-only state was\naccepted in the probe); from the SDK, pass an empty string as state with\nimages. The CLI always needs a state.\n\nThe rules:\nUp to 8 images per request. Measured: 8 images with three questions were\n  accepted, and 9 images with one question were refused with a 400 whose message\n  words the cap as \"per question\". NeuroLink applies the 8 to the request as a\n  whole.\nPNG, JPEG or WebP only. The type comes from the bytes, not the file\n  extension. A GIF, or a data: URL of another type, is refused before any\n  request, as a non-retryable invalid_request. The API refuses a GIF with a 400\n  too.\nEach is a Buffer, a local file path or a data: URL. An http(s) URL is\n  refused locally and never fetched. The API does not fetch URLs either: an\n  https image URL returned a 400 in the probe.\nEach image may be at most 2,048 tiles of 32 × 32 pixels. Round the width\n  and the height to the nearest multiple of 32 and keep (width / 32) ×\n  (height / 32) at or under 2,048: by that count 1440 × 1440 (2,025 tiles) and\n  2048 × 1024 (2,048) fit, and 1600 × 1310 (2,050) does not. Perplexity documents\n  this rule, and it was checked at its edge. Fitting sizes were answered in\n  about a second: 1920 × 1080 (2,040 tiles) in 1.1 seconds, 2048 × 1024 (2,048,\n  exactly the cap) in 1.3 and 1450 × 1450 in 1.1.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"Images","lvl3":""}},
-{"objectID":"8ad813ca01099f441bbc2db40ee064d9a0294d269166c42a6f87fce2a1c1330f","title":"When NeuroLink uses it","url":"/docs/getting-started/providers/perplexity-decider#when-neurolink-uses-it","content":"Every built-in consumer of decide asks for the default decision provider. That\nis the first one that is configured, in the environment or in the credentials\npassed to the SDK, in the order TypeSafe, Laya, XOR, SubQ,\nPerplexity,\nCloudflare Clef, then\nOpenAI Decisions. TypeSafe counts\nwith either of its keys, TYPESAFE_API_KEY or AI_GATEWAY_API_KEY. Laya and XOR\neach count only with both their key and their base URL. SubQ counts with its key\nalone, SUBQ_API_KEY. Perplexity counts with its key alone, a non-blank\nPERPLEXITY_API_KEY or credentials.perplexityDecider.apiKey. Cloudflare Clef\ncounts only with both CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID. OpenAI Decisions counts with\nits key alone, OPENAI_API_KEY or credentials.openaiDecisions.apiKey. A caller\ncan always name Perplexity with provider: \"perplexity-decider\". So:\nOnly a Perplexity key: built-in features use Perplexity.\nA Perplexity key, plus TypeSafe's key, or Laya's or XOR's key and base\n  URL, or SUBQ_API_KEY: built-in features use TypeSafe, Laya, XOR or SubQ,\n  in that order, whichever is configured. Perplexity runs only where a caller\n  asks for provider: \"perplexity-decider\".\nA Perplexity key, plus CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID:\n  built-in features use Perplexity; Cloudflare Clef runs only where a caller\n  asks for provider: \"cloudflare-clef\".\nA key for the text provider only: the same as the first case. The key is\n  shared, so the text provider's key also configures this one. See\n  One key, two providers.\nNone of TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef or OpenAI Decisions\n  configured: everything behaves exactly as it did without a decision model.\n\nThe classifier router's auto gate counts the credentials given to the\nNeuroLink constructor, the environment and the request's own per-call\ncredentials, so a per-call credentials.perplexityDecider can make auto pick\njev.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"When NeuroLink uses it","lvl3":""}},
+{"objectID":"e1b9161659de68a10277e69ce53077f2aed241892d783054f01471fb793024bd","title":"Images","url":"/docs/getting-started/providers/perplexity-decider#images","content":"Perplexity reads images alongside the state. TypeSafe, Laya,\nSubQ and ScaleDown do not read media. XOR reads images and a video; Perplexity reads images only, and so do\nCloudflare Clef, up to 4 of them, and\nOpenAI Decisions. In the probe of\nPerplexity's model it did read an image: asked red versus blue about one, it\nanswered correctly at about 0.98 confidence.\n\nFrom the CLI, pass --image <path>, repeated for several:\n\nOn the wire the images go inside state: NeuroLink sends state as an array,\nyour state first and then one OpenAI-style image_url part per image, each a\nbase64 data URL. The API has no separate images field and rejects unknown\nfields. An image can also be the whole input (an image-only state was\naccepted in the probe); from the SDK, pass an empty string as state with\nimages. The CLI always needs a state.\n\nThe rules:\nUp to 8 images per request. Measured: 8 images with three questions were\n  accepted, and 9 images with one question were refused with a 400 whose message\n  words the cap as \"per question\". NeuroLink applies the 8 to the request as a\n  whole.\nPNG, JPEG or WebP only. The type comes from the bytes, not the file\n  extension. A GIF, or a data: URL of another type, is refused before any\n  request, as a non-retryable invalid_request. The API refuses a GIF with a 400\n  too.\nEach is a Buffer, a local file path or a data: URL. An http(s) URL is\n  refused locally and never fetched. The API does not fetch URLs either: an\n  https image URL returned a 400 in the probe.\nEach image may be at most 2,048 tiles of 32 × 32 pixels. Round the width\n  and the height to the nearest multiple of 32 and keep (width / 32) ×\n  (height / 32) at or under 2,048: by that count 1440 × 1440 (2,025 tiles) and\n  2048 × 1024 (2,048) fit, and 1600 × 1310 (2,050) does not. Perplexity documents\n  this rule, and it was checked at its edge.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"Images","lvl3":""}},
+{"objectID":"8ad813ca01099f441bbc2db40ee064d9a0294d269166c42a6f87fce2a1c1330f","title":"When NeuroLink uses it","url":"/docs/getting-started/providers/perplexity-decider#when-neurolink-uses-it","content":"Every built-in consumer of decide asks for the default decision provider. That\nis the first one that is configured, in the environment or in the credentials\npassed to the SDK, in the order TypeSafe, Laya, XOR, SubQ,\nPerplexity,\nCloudflare Clef,\nOpenAI Decisions, then ScaleDown. TypeSafe counts\nwith either of its keys, TYPESAFE_API_KEY or AI_GATEWAY_API_KEY. Laya and XOR\neach count only with both their key and their base URL. SubQ counts with its key\nalone, SUBQ_API_KEY. Perplexity counts with its key alone, a non-blank\nPERPLEXITY_API_KEY or credentials.perplexityDecider.apiKey. Cloudflare Clef\ncounts only with both CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID. OpenAI Decisions counts with\nits key alone, OPENAI_API_KEY or credentials.openaiDecisions.apiKey.\nScaleDown counts with its key alone, SCALEDOWN_API_KEY or\ncredentials.scaledown.apiKey, which no other provider reads. A caller\ncan always name Perplexity with provider: \"perplexity-decider\". So:\nOnly a Perplexity key: built-in features use Perplexity.\nA Perplexity key, plus TypeSafe's key, or Laya's or XOR's key and base\n  URL, or SUBQ_API_KEY: built-in features use TypeSafe, Laya, XOR or SubQ,\n  in that order, whichever is configured. Perplexity runs only where a caller\n  asks for provider: \"perplexity-decider\".\nA Perplexity key, plus CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID:\n  built-in features use Perplexity; Cloudflare Clef runs only where a caller\n  asks for provider: \"cloudflare-clef\".\nA Perplexity key, plus SCALEDOWN_API_KEY: built-in features use\n  Perplexity, which comes ahead of ScaleDown; ScaleDown runs only where a caller\n  asks for provider: \"scaledown\".\nA key for the text provider only: the same as the first case. The key is\n  shared, so the text provider's key also configures this one. See\n  One key, two providers.\nNone of TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef, OpenAI Decisions\n  or ScaleDown configured: everything behaves exactly as it did without a decision model.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"When NeuroLink uses it","lvl3":""}},
 {"objectID":"9545ac1f3da9bfc59774adf1978baa598ba4fc65670b5d0d4c96f7c10e2a1c3c","title":"What is sent to Perplexity","url":"/docs/getting-started/providers/perplexity-decider#what-is-sent-to-perplexity","content":"Perplexity receives the state and the questions of every decision made on\nyour behalf. The questions are NeuroLink's fixed wording, plus text from your own\nconversation where a consumer quotes it; the state is where your text goes.\nThis is what each built-in consumer sends and what turns it on, read from the\ncode:\nModel routing. One request behind the router, the model catalogue and the\n  per-request context budget.\nSends: the first 8,000 characters of the prompt of the call being routed,\n    plus whether tools are available, whether the request includes images, the\n    estimated input tokens, the requested thinking level, whether a session is\n    bound and how many prior messages the caller passed in. When the pool has\n    more than one model, the questions carry the ids and descriptions of the\n    models in it. The images, the session id and the conversation itself are not\n    sent.\nTurned on by: classifierRouter.enabled with a pool: an opt-in. With\n    the default classifier: \"auto\", a configured decision provider selects the\n    decision strategy. classifier: \"heuristic\" or \"llm\" never calls a\n    decision provider. A call that pins both provider and model is not\n    routed.\nRelevance compaction. Stage 0 of context compaction.\nSends: the current request (first 4,000 characters) and, for each earlier\n    user or assistant message that is plain text (not a tool call or result, a\n    summary or a pinned skill), its role and its first 1,200 characters. That\n    text travels twice, once in the state and once in the message's question. By\n    default the six most recent messages are never sent\n    (contextRelevance.protectRecent), and at most 300 are asked about.\nTurned on by: nothing beyond a configured decision provider. It runs\n    inside context compaction, which starts when a conversation outgrows its\n    context budget.\nSummary-quality gate. Guards the summarization stage of compaction.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"What is sent to Perplexity","lvl3":""}},
 {"objectID":"01342235e59d6796c886025eea6c3dbdefc5f4420f47123db9b23eacc22955d3","title":"One key, two providers","url":"/docs/getting-started/providers/perplexity-decider#one-key-two-providers","content":"PERPLEXITY_API_KEY is the variable the\nPerplexity text provider reads to serve generate() and\nstream(). It is also the variable that configures this provider. Nothing else\nneeds to be set, so a host that set the key only to use Sonar has also\nconfigured decide:\nWith none of TypeSafe, Laya, XOR or SubQ configured, the texts listed in\n  What is sent to Perplexity go to Perplexity's\n  Decisions API. Context compaction needs no other opt-in, so a host with long\n  conversations starts sending conversation text the first time one outgrows its\n  budget. Model routing, tool routing and RAG planning do so only where you have\n  turned them on. No code changes: the key is the switch.\nIt is fail-open, and TypeSafe, Laya, XOR and SubQ take precedence. If you have\n  configured any of them, the shared key changes nothing for the built-in\n  consumers, and a Perplexity failure leaves behaviour as it was.\nWith no decision provider configured at all, nothing is sent.\n\nAn unedited copy of .env.example does not set the key, because its\nPERPLEXITY_API_KEY line is commented out.\n\nA key in a .env file counts as a key in the environment. Importing the SDK\nloads the .env in the current working directory, and the CLI does the same at\nstart; when DOTENV_CONFIG_PATH is set, that file is loaded instead. A key that\nsits in .env therefore configures decide exactly as one exported in the\nshell does.\n\nTo keep a key that was set for Sonar from activating decisions, use one of\nthese:\nHand the key to the text provider through the SDK, not the environment, and\n   keep it out of .env.\n   new NeuroLink({ credentials: { perplexity: { apiKey } } }) configures the\n   text provider and not this one, because NeuroLink looks for this provider's\n   key in PERPLEXITY_API_KEY and in credentials.perplexityDecider, never in\n   credentials.perplexity.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"One key, two providers","lvl3":""}},
 {"objectID":"2c9da95e9403e04ac92a64727b9b098a3b1e626742b59f4ffebc6e24914ec316","title":"Measured on a real account, October 2026","url":"/docs/getting-started/providers/perplexity-decider#measured-on-a-real-account-october-2026","content":"NeuroLink probed the live API from an account limited to 10 requests per second.\nThese figures were observed, not taken from Perplexity's documentation:\n\n| What                  | Measured                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |\n| --------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |\n| Questions per request | 128 accepted (128 answers in 8.9 s, 11,904 input tokens); 129 is refused with a 400. An extra minimal question adds about 92 input tokens, and a one-character state with one question bills 116.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"Measured on a real account, October 2026","lvl3":""}},
@@ -6999,7 +7004,7 @@
 {"objectID":"e0c0e895220d617b9c97e379206986672c638c01de25fb9e0b52fa7c04bf0ca1","title":"Rate limits","url":"/docs/getting-started/providers/perplexity-decider#rate-limits","content":"The account measured allows 10 requests per second. A 429 is retried once, after\nthe wait its Retry-After header asks for. NeuroLink honours that header on any\nretried reply that carries it, and only as a whole number of seconds above zero\nor as an HTTP date in the future; anything else (zero, a negative number, a hex\nor exponent spelling, a fraction, a date already past) is ignored, and the\ndefault backoff of a quarter to half a second is used instead. A wait longer\nthan what is left of the attempt's timeout is not taken: the error is thrown at\nonce so a fail-open consumer is not held for the length of a cooldown, and a\ncaller's abort signal ends a wait immediately.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"Rate limits","lvl3":""}},
 {"objectID":"f40b2da5feaf02bc4eaf5102d75c33ce6fc9f979f7106853002b142aad4fd67f","title":"Errors","url":"/docs/getting-started/providers/perplexity-decider#errors","content":"| Reply                                                                | Kind                                                    | Retried                        |\n| -------------------------------------------------------------------- | ------------------------------------------------------- | ------------------------------ |\n| 401                                                                  | authentication — the provider instance stops retrying | no                             |\n| 413                                                                  | max_tokens_exceeded (the body is over 32 MiB)         | no                             |\n| 400 saying the input exceeds the model's maximum context length      | max_tokens_exceeded                                   | no                             |\n| any other 4xx (for example another 400, 403, 404)                    | invalid_request, with Perplexity's own message        | no                             |\n| 429                                                                  | rate_limit                                            | yes, once, after Retry-After |\n| 503                                                                  | overloaded                                            | yes, once                      |\n| other 5xx (500, 502, 504)                                            | server                                                | yes, once                      |\n| no response within the timeout                                       | timeout                                               | yes, once                      |\n| the request fails before any reply (connection refused, DNS failure) | network                                               | yes, once                      |\n| the caller's own signal aborts the request                         | network                                               | no                             |\n| state or questions over the limit (local)","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"Errors","lvl3":""}},
 {"objectID":"a812289efc882316bb62c31aea533cd0a0e8e219eeea12289b4503450b4dd9a5","title":"Troubleshooting","url":"/docs/getting-started/providers/perplexity-decider#troubleshooting","content":"Perplexity requires an API key — set PERPLEXITY_API_KEY, or pass\n  credentials.perplexityDecider.apiKey. A key passed as\n  credentials.perplexity does not count.\nA 401, Invalid API key provided — Perplexity rejected the key, and it\n  checks the key before anything else. The provider instance stops retrying\n  after a rejection, so fix the key and construct a new one.\nA 400 naming the model — PERPLEXITY_DECIDER_MODEL, or a per-call\n  model, names something other than pplx-decider-v1-27b.\nThe Perplexity base URL must not carry credentials… — the override has a\n  user name, a password, a query string or a fragment. Set it to the origin only,\n  or leave it unset for https://api.perplexity.ai.\nThe Perplexity base URL must start with https:// or http:// — the value\n  parses as a URL with another scheme: an explicit one such as ftp://, or a host\n  name and port with no scheme, such as localhost:8080, which reads as the\n  scheme localhost:. Put https:// in front, or http:// for a server on this\n  machine.\nThe Perplexity base URL must use https:// unless it points at this machine —\n  the override is an http:// URL whose host is not localhost, 127.0.0.1 or\n  [::1]. A plain-http request would send the key unencrypted. Use https://, or\n  reach the server through a port-forward to localhost.\nPerplexity answered with a redirect — the endpoint replied 301, 302, 307 or 308. The provider does not follow it and does not retry. Set the base URL to the\n  address the redirect leads to.\nThe Perplexity base URL is not a valid absolute URL — the value does not\n  parse as a URL at all: a bare host name such as api.perplexity.ai (the likely\n  mistake), an IP address and port such as 127.0.0.1:8080, a path, or a scheme\n  with no host. Write the origin with its scheme.\nA refused base URL in the debug log — it is not there.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"Troubleshooting","lvl3":""}},
-{"objectID":"6d580e7ffb3867905a72879cdccbab76b5df280677e9d18b0029fd5bc5092025","title":"See also","url":"/docs/getting-started/providers/perplexity-decider#see-also","content":"The decide inference type\nTypeSafe (Jev) Provider Guide\nLaya Provider Guide\nXOR Provider Guide\nCloudflare Clef Provider Guide\nSubQ Provider Guide\nOpenAI Decisions Provider Guide\nPerplexity text provider\nPerplexity Decisions API reference","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"See also","lvl3":""}},
+{"objectID":"6d580e7ffb3867905a72879cdccbab76b5df280677e9d18b0029fd5bc5092025","title":"See also","url":"/docs/getting-started/providers/perplexity-decider#see-also","content":"The decide inference type\nTypeSafe (Jev) Provider Guide\nLaya Provider Guide\nXOR Provider Guide\nCloudflare Clef Provider Guide\nSubQ Provider Guide\nOpenAI Decisions Provider Guide\nScaleDown Provider Guide\nPerplexity text provider\nPerplexity Decisions API reference","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Decisions Provider Guide","lvl2":"See also","lvl3":""}},
 {"objectID":"6d55ee7ec0f21369a70b421eaeaddf6b44fe88746beb28c252588d42d97048b7","title":"Perplexity Provider Guide","url":"/docs/getting-started/providers/perplexity","content":"Perplexity Provider Guide\n\nWeb-search-augmented generation via Perplexity's Sonar models\n\nOverview\n\nPerplexity hosts a family of LLMs (Sonar,\nSonar Pro, Sonar Reasoning) that pair the model with a live web search\nbackend; responses include citations to the documents the model relied on.\n\nKey Facts\nProtocol: OpenAI-compatible (/chat/completions)\nDefault base URL: https://api.perplexity.ai\nDefault model: sonar\nCitations: Returned via citations field on the response\nStreaming: Yes\nTool calling: No\n\nQuick Start\nGet an API Key\n\nhttps://www.perplexity.ai/settings/api\nConfigure\n\nPERPLEXITY_API_KEY also configures decisions. The same variable configures\nthe Perplexity Decisions provider (perplexity-decider,\nwhich serves decide()). So setting it for Sonar also lets NeuroLink's built-in\nfeatures send decision requests to Perplexity when no TypeSafe, Laya, XOR or SubQ\nprovider is configured, and context compaction needs no other opt-in to do so.\nTo avoid that, pass the key through the SDK as credentials.perplexity.apiKey\ninstead of the environment, and keep it out of .env too: the SDK and the CLI\nload that file into the environment, so a PERPLEXITY_API_KEY line like the one\nabove would still configure decisions. (That the text provider reads this slice\nis taken from the code and has not been run against the live API.) See\nWhat is sent to Perplexity\nand One key, two providers.\nGenerate\n\nSupported Models\n\n| Model ID          | Notes                               |\n| ----------------- | ----------------------------------- |\n| sonar           | Default; fast search-augmented chat |\n| sonar-pro       | Larger context, deeper search       |\n| sonar-reasoning | Chain-of-thought reasoning          |\n\nCLI Usage\n\nProvider Aliases\n\n| Alias        | Example                 |\n| ------------ | ----------------------- |\n| perplexity | --provider perplexity |\n\nConfiguration Reference\n\n| Environment Variable  | Required | Default                     |\n| --------------------- | -------- | --------------------------- |\n| PERPLEXITY_API_KEY  | Yes      | —                           |\n| PERPLEXITY_MODEL    | No       | sonar                     |\n| PERPLEXITY_BASE_URL | No       | https://api.perplexity.ai |\n\nFeature Support Matrix\n\n| Feature           | Support      |\n| ----------------- | ------------ |\n| Text generation   | Yes          |\n| Streaming         | Yes          |\n| Tool calling      | No           |\n| Structured output | Limited      |\n| Web search        | Yes (native) |\n| Citations         | Yes          |\n\nSee Also\nPerplexity Decisions Provider — typed, calibrated decide() judgments on the same key\nAnthropic Provider — web-search via the web_search tool\nVertex Provider — Google search grounding","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"59d597c2561bea4977d9a4ae9a111235c2a042b1573e0fad7cf2147851b79e13","title":"Perplexity Provider Guide","url":"/docs/getting-started/providers/perplexity#perplexity-provider-guide","content":"Web-search-augmented generation via Perplexity's Sonar models","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Provider Guide","lvl2":"Perplexity Provider Guide","lvl3":""}},
 {"objectID":"755494265419901fd106a26a27844178614744eacb256af720b7acfdc700a22c","title":"Overview","url":"/docs/getting-started/providers/perplexity#overview","content":"Perplexity hosts a family of LLMs (Sonar,\nSonar Pro, Sonar Reasoning) that pair the model with a live web search\nbackend; responses include citations to the documents the model relied on.","hierarchy":{"lvl0":"Getting Started","lvl1":"Perplexity Provider Guide","lvl2":"Overview","lvl3":""}},
@@ -7176,6 +7181,29 @@
 {"objectID":"cd1de30168eb04f05817488311cca7b4f12bf8367f224d54df7d0b972af81d3e","title":"Verification status","url":"/docs/getting-started/providers/sarvam#verification-status","content":"Tier-2 onboarding requires evidence before a provider is accepted, and\npnpm run verify:provider-onboarding gates it in CI. This is what the\ncatalog records for Sarvam AI — and, unusually, all of it is credential-free:\n\n| Probe                 | Result                                                                                                                                                                                                                                                                                    |\n| --------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Roster                | Unauthenticated GET https://api.sarvam.ai/v1/models, HTTP 200, 2026-09-28 — returned exactly sarvam-105b and sarvam-105b-conversations, no API key used.                                                                                                                        |\n| Auth rejection        | Not probed live (would require a POST request, out of scope for credential-free onboarding). The 403 shape below is transcribed from Sarvam's own authentication docs, not observed.                                                                                                      |\n| Wire compatibility    | Documentary only — Sarvam's API reference explicitly describes an OpenAI-shaped messages / choices / finish_reason / object: \"chat.completion\" schema, tools/tool_choice, response_format (json_schema, json_object), and stream: true (SSE) for this endpoint.       |\n| Live capability sweep | Not run. evidence.liveMatrix is null.","hierarchy":{"lvl0":"Getting Started","lvl1":"Sarvam AI Provider Guide","lvl2":"Verification status","lvl3":""}},
 {"objectID":"8e57df7483555e36212512809e787dc851cae1ddb71ca3758faab8f991880cad","title":"Troubleshooting","url":"/docs/getting-started/providers/sarvam#troubleshooting","content":"| Symptom                                    | Cause                                                                                | Fix                                                                              |\n| ------------------------------------------ | ------------------------------------------------------------------------------------ | -------------------------------------------------------------------------------- |\n| Invalid Sarvam AI API key                | SARVAM_API_KEY unset or wrong                                                      | Check the key at https://dashboard.sarvam.ai                                     |\n| HTTP 403 on every call                     | Sarvam returns 403 (not 401) for both bad and forbidden keys                         | Confirm the key is current and has access; see error.code in the response      |\n| Model not found                            | Only sarvam-105b and sarvam-105b-conversations exist                             | Use one of those two ids — the endpoint serves no others                         |\n| Empty/short replies at a small maxTokens | reasoning_effort defaults to medium, and reasoning_content is billed as output | Give reasoning prompts a generous maxTokens budget or lower reasoning_effort |","hierarchy":{"lvl0":"Getting Started","lvl1":"Sarvam AI Provider Guide","lvl2":"Troubleshooting","lvl3":""}},
 {"objectID":"49895306150970c4315329ad7d39d3141c62342bd1ca0ad4643cc8c3b63b1799","title":"See also","url":"/docs/getting-started/providers/sarvam#see-also","content":"Provider setup overview\nAll providers\nTier-2 onboarding — how this provider's JSON becomes a working integration\nProvider feature compatibility","hierarchy":{"lvl0":"Getting Started","lvl1":"Sarvam AI Provider Guide","lvl2":"See also","lvl3":""}},
+{"objectID":"a5f27aad5d289af1b5605c874f662cf8feb10a9311baaef5577fa0788921a31c","title":"ScaleDown Provider Guide","url":"/docs/getting-started/providers/scaledown","content":"ScaleDown Provider Guide\n\nA provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR,\nSubQ, Perplexity,\nCloudflare Clef and\nOpenAI Decisions, from ScaleDown's hosted Classify\nmodels. It emits no text at all.\n\nWhere the facts about ScaleDown come from. Everything this guide says about\nScaleDown's endpoint comes from ScaleDown's own documentation and from probes\nthat the marketing session made on 2026-10-15. NeuroLink's project did not\nprobe it: no request to ScaleDown was made in building this provider, and the\n200 replies in NeuroLink's tests are stand-ins written from the documented\nshape, and the error texts in the mocked tests are the ones the marketing\nsession reported.\nEach item that nobody measured is listed under\nNot measured.\n\nOverview\n\nclassify-1 is ScaleDown's decision model. You send one state plus named,\ntyped questions. The model answers every question in a single request and\nreturns a typed answer for each, with a probability instead of a sentence.\nScaleDown speaks the same System One wire as TypeSafe, so NeuroLink describes it\nwith one catalog file, src/lib/providers/catalog/scaledown.json, and no\nprovider code of its own. NeuroLink calls ScaleDown's public endpoint, so a key\nalone configures it: there is no base URL to set.\n\nKey Facts\n\n|                       |                                                                                                                                                                                     |\n| --------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Provider id           | scaledown; the alias scaledown-classify names it too                                                                                                                            |\n| Inference type        | decide only                                                                                                                                                                       |\n| Models                | classify-1 (the default) and classify, which ScaleDown describes as its latest Classify model. Any other name is refused by ScaleDown with HTTP 422. SCALEDOWN_MODEL sets one |\n| Key                   | SCALEDOWN_API_KEY, or credentials.scaledown.apiKey. The variable is ScaleDown's own: no text provider reads it                                                                  |\n| Key header            | x-api-key: <key>. NeuroLink never sends Authorization: Bearer to ScaleDown                                                                                                      |\n| Endpoint              | POST https://api.scaledown.xyz/v1/scaledown; SCALEDOWN_BASE_URL or credentials.scaledown.baseURL can name another base URL (see Base URL)                        |\n| State                 | text only: a string goes out unchanged, any other value as its JSON text; see How the state is sent                                                       |\n| Questions per request | no cap declared                                                                                                                                                                     |\n| State window          | none declared, so nothing is refused locally on size; ScaleDown's input ceiling is not measured                                                                                     |\n| Media                 | none. Images and video attached to a decide() call are refused before any request; see Images and video                                                      |\n| Default timeout       | 5 seconds per attempt; timeoutMs or --timeout replaces it; see Timeouts                                                                                            |\n| Cost                  | priced per input token, input only; see Cost for the rate, its source and its date                                                                                         |\n| Precedence            | the last of the eight: used automatically only when none of TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef or OpenAI Decisions is configured                                |\n| Status                | NeuroLink lists both models as preview                                                                                                                                            |\n\nQuick Start\nGet a key\n\nCreate an account and an API key by following\nScaleDown's documentation. Sign-up gives a free\nallowance; its expiry is not established. ScaleDown's documentation says a\nrequest made with no credit left returns HTTP 402.\nConfigure\n\nSet the key in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:\n\ncredentials.scaledown is its own","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"","lvl3":""}},
+{"objectID":"b9df3945bc87bb36e60b7b96008ffdf5fa0d81362bb0d14c8be9be28c055b127","title":"ScaleDown Provider Guide","url":"/docs/getting-started/providers/scaledown#scaledown-provider-guide","content":"A provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR,\nSubQ, Perplexity,\nCloudflare Clef and\nOpenAI Decisions, from ScaleDown's hosted Classify\nmodels. It emits no text at all.\n\nWhere the facts about ScaleDown come from. Everything this guide says about\nScaleDown's endpoint comes from ScaleDown's own documentation and from probes\nthat the marketing session made on 2026-10-15. NeuroLink's project did not\nprobe it: no request to ScaleDown was made in building this provider, and the\n200 replies in NeuroLink's tests are stand-ins written from the documented\nshape, and the error texts in the mocked tests are the ones the marketing\nsession reported.\nEach item that nobody measured is listed under\nNot measured.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"ScaleDown Provider Guide","lvl3":""}},
+{"objectID":"5cd76a87b19e0ac4507befcfdbf6fe2a0b60672ef4b39ba8a48d7108d7e03641","title":"Overview","url":"/docs/getting-started/providers/scaledown#overview","content":"classify-1 is ScaleDown's decision model. You send one state plus named,\ntyped questions. The model answers every question in a single request and\nreturns a typed answer for each, with a probability instead of a sentence.\nScaleDown speaks the same System One wire as TypeSafe, so NeuroLink describes it\nwith one catalog file, src/lib/providers/catalog/scaledown.json, and no\nprovider code of its own. NeuroLink calls ScaleDown's public endpoint, so a key\nalone configures it: there is no base URL to set.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"Overview","lvl3":""}},
+{"objectID":"1dc9cfb0502df137f78e02d2f0ce01f1c20021432f32b55e457f4af9aa8c28d3","title":"Key Facts","url":"/docs/getting-started/providers/scaledown#key-facts","content":"|                       |                                                                                                                                                                                     |\n| --------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Provider id           | scaledown; the alias scaledown-classify names it too                                                                                                                            |\n| Inference type        | decide only                                                                                                                                                                       |\n| Models                | classify-1 (the default) and classify, which ScaleDown describes as its latest Classify model. Any other name is refused by ScaleDown with HTTP 422. SCALEDOWN_MODEL sets one |\n| Key                   | SCALEDOWN_API_KEY, or credentials.scaledown.apiKey. The variable is ScaleDown's own: no text provider reads it                                                                  |\n| Key header            | x-api-key: <key>.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"Key Facts","lvl3":""}},
+{"objectID":"6e1271629d7848d082fcee808db1251840c95379629429d3af24aaacf4cbb453","title":"1. Get a key","url":"/docs/getting-started/providers/scaledown#1-get-a-key","content":"Create an account and an API key by following\nScaleDown's documentation. Sign-up gives a free\nallowance; its expiry is not established. ScaleDown's documentation says a\nrequest made with no credit left returns HTTP 402.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"1. Get a key","lvl3":""}},
+{"objectID":"59465c6b09139d6862d9a9a23650fd4ef6e24a036f6b80fd29b066d42d2044d2","title":"2. Configure","url":"/docs/getting-started/providers/scaledown#2-configure","content":"Set the key in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:\n\ncredentials.scaledown is its own slice (apiKey and baseURL). A blank\nSCALEDOWN_API_KEY or credentials.scaledown.apiKey counts as no key.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"2. Configure","lvl3":""}},
+{"objectID":"55f2cd0ebb6971a4211f7f9d1221cf78ea3a539224d700f623346f317039322e","title":"Base URL","url":"/docs/getting-started/providers/scaledown#base-url","content":"SCALEDOWN_BASE_URL and credentials.scaledown.baseURL name the part of the URL\nin front of /v1/scaledown. It can be an origin, such as\nhttps://api.scaledown.xyz, or an origin with a path prefix, such as a gateway\nthat forwards https://gateway.example.com/scaledown to ScaleDown. NeuroLink\ndrops trailing slashes and appends /v1/scaledown, so a value that already ends\nin /v1 would send the request to /v1/v1/scaledown. A blank value is no\noverride. NeuroLink checks nothing else about the value before the request.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"Base URL","lvl3":""}},
+{"objectID":"7d53a332341c36ef5f7392a98a61ec8bbf47e87a5d57ec9f47ad8ffdfdc79a18","title":"3. Use it","url":"/docs/getting-started/providers/scaledown#3-use-it","content":"result.model is the model ScaleDown's reply names, which can differ from the\none you asked for (classify is ScaleDown's name for its latest Classify model).\nresult.usage.inputTokens is the reply's usage.input_tokens.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"3. Use it","lvl3":""}},
+{"objectID":"6fa47bf20bb49567239e1e7aa2aaf60c0ddf14f94b5ecd5c9291325af97ee1cc","title":"From the CLI","url":"/docs/getting-started/providers/scaledown#from-the-cli","content":"The text output ends with a Cost: line, priced from the input tokens the reply\nreports (see Cost).","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"From the CLI","lvl3":""}},
+{"objectID":"f0a725ba02edca902073a9425b5e6f11d56bb918130f79c1f4f75c017fb7ab37","title":"How a request is sent","url":"/docs/getting-started/providers/scaledown#how-a-request-is-sent","content":"ScaleDown's wire is System One's, so you write the same questions as for any\nother decide provider.\n\n| NeuroLink | Sent to /v1/scaledown                                                                                    |\n| --------- | ---------------------------------------------------------------------------------------------------------- |\n| the key   | the header x-api-key: <key>, beside Content-Type: application/json; there is no Authorization header |\n| the model | model in the body (classify-1 unless you set another)                                                  |\n| state   | state, always a string; see below                                                                        |\n| boolean | a noul question with its instructions                                                                  |\n| choice  | a choice question; criteria maps each option name to its description                                   |\n| score   | a score question; criteria is an ordered list of level descriptions                                    |\n\nThe reply has the System One shape, an answers object keyed by question id,\nplus model and usage (input_tokens, output_tokens and a cost in USD).\nA boolean answer carries a probability. A choice or score answer carries\nthe confidence the reply reports and a probability for each option or level; when\na reply omits the confidence, NeuroLink uses the highest probability in that\nanswer's distribution instead. Whether ScaleDown's confidence is calibrated is\nnot established, so tune a threshold on it against your own labelled data.\n\nScaleDown's documentation describes a score question's criteria as an ordered\nlist of 2 to 10 level descriptions. NeuroLink does not enforce that.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"How a request is sent","lvl3":""}},
+{"objectID":"019697cf87430bd6504ed998cc379cd4c9565713e8336341735eac9e28a05e95","title":"How the state is sent","url":"/docs/getting-started/providers/scaledown#how-the-state-is-sent","content":"ScaleDown's documentation says it reads only the text and document fields of\nthe state, and treats a string state as shorthand for { \"text\": <the string> }.\nNeuroLink therefore sends the state as a string, whatever you pass:\n\n| You pass                                  | ScaleDown receives                              |\n| ----------------------------------------- | ----------------------------------------------- |\n| a string                                  | the string, unchanged (not trimmed)             |\n| an object, an array, a number, a boolean  | its JSON text, as one string (JSON.stringify) |\n| {} or []                              | the two-character text {} or []             |\n| a string that is empty or only whitespace | nothing: refused locally, before any request    |\n\nTwo consequences:\nAn object state is classified as text. ScaleDown does not read an object's\n  other keys, so { \"title\": \"...\", \"review\": \"...\" } reaches it as the JSON text\n  of that object, and an object that has a text field is not unwrapped either.\n  If you want it to read plain prose, pass a string.\n{} and [] are not refused. They are non-blank text, so they are sent\n  and ScaleDown classifies the literal characters. ScaleDown itself answers HTTP\n  422 ('state.text' or 'state.document' must be provided) to an empty or\n  text-less state object, as the marketing session saw on 2026-10-15, but\n  NeuroLink never sends one. Check for an empty container yourself if you want it\n  refused.\n\nA blank string is refused as an invalid_request, with a message that names\nScaleDown and the rule (needs non-blank state text) and never repeats the state.\ndecide() rejects with it, tryDecide() returns null, and no request is made.\nThis is a catalog setting, decide.stateEncoding: see\nHow the catalog entry says this.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"How the state is sent","lvl3":""}},
+{"objectID":"909b8119299644649ccc89ea4d6942c9d088298200c8dde9487649c6a689140f","title":"Choice options and billing","url":"/docs/getting-started/providers/scaledown#choice-options-and-billing","content":"ScaleDown's documentation says every option of a choice question is scored by\na separate internal call, so the input tokens a request is charged for grow with\nthe number of options times the length of the state. The marketing session\nreported two data points on 2026-10-15, not measured here: about 1,829 input\ntokens for a 20-option choice and about 13,059 for a 255-option choice, each on a\none-sentence state.\n\nNeuroLink declares no cap on options. A probe by the marketing session saw a\n255-option choice answered and a 300-option choice refused with HTTP 502; the\nlimit between them is not mapped. When ScaleDown answers a request with HTTP 502,\nas it did to the 300-option probe, NeuroLink reports a server error, retries\nonce, and every built-in consumer then treats the decision as failed and falls\nback. Whether a refused request is charged is not established. Keep the number of\noptions of a choice question modest, and see Cost.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"Choice options and billing","lvl3":""}},
+{"objectID":"0e9ffbaad88dadb6493a28fd7481227ff45f814672b9224e965f48bf3b8b9e0b","title":"When NeuroLink uses it","url":"/docs/getting-started/providers/scaledown#when-neurolink-uses-it","content":"Every built-in consumer of decide — model routing, relevance-driven\ncompaction, tool routing and RAG planning — asks for the default decision\nprovider. That is the first one that is configured, in the environment or in\nthe credentials passed to the SDK, in the order TypeSafe, Laya,\nXOR, SubQ, Perplexity,\nCloudflare Clef, OpenAI Decisions,\nthen ScaleDown. ScaleDown counts with its key alone. So:\nOnly a ScaleDown key: built-in features use ScaleDown, and each of their\n  decisions is then a paid request (see Cost) that still falls back to\n  what NeuroLink did before on any failure.\nA ScaleDown key, plus any earlier provider's key: built-in features use\n  that earlier provider. ScaleDown runs only where a caller asks for\n  provider: \"scaledown\".\nA blank SCALEDOWN_API_KEY: no key. ScaleDown is not configured.\nprovider: \"scaledown\" with no key: an authentication error and no\n  request.\n\nScaleDown is last on purpose, and its key is not shared. Apart from ScaleDown, the providers that\ncome after SubQ in the order are the ones whose key is also read by a text\nprovider (Perplexity's, the Cloudflare pair and OpenAI's), placed behind it so\nthat a key set only for a text provider cannot displace a decision provider a\nhost set up on purpose. That reasoning does not cover ScaleDown, whose key\nbelongs to no other provider: it was placed last, after OpenAI Decisions, as a\ndeliberate choice and not by that rule. The effect is that an ambient OPENAI_API_KEY, PERPLEXITY_API_KEY or\nCloudflare pair in the environment outranks a SCALEDOWN_API_KEY set on purpose.\nA host that wants ScaleDown for the built-in features must not have the earlier\nproviders configured, or must name it in its own calls.\n\nThe classifier router's auto strategy counts SCALEDOWN_API_KEY too: with\nonly that key set, auto resolves to jev and ScaleDown answers.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"When NeuroLink uses it","lvl3":""}},
+{"objectID":"b06a4af3a93c004935822bcc4b8c3fa7b19a496460ca3d2a9f8ea72446d6c262","title":"How the catalog entry says this","url":"/docs/getting-started/providers/scaledown#how-the-catalog-entry-says-this","content":"scaledown.json needed two optional keys in its decide block that no earlier\nentry set, both read by the shared catalog engine for any entry that sets them.\nAn entry that omits them behaves exactly as before.\n\n| Key             | Value         | What it does                                                                                                                                                                                                                                                                                   |\n| --------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| authHeader    | \"x-api-key\" | The API key is sent as x-api-key: <key> and no Authorization header is sent at all. \"bearer\", the default when the key is omitted, sends Authorization: Bearer <key>. This is the decide request's own setting; quirks.authHeaderStyle belongs to the chat path and is not read here |\n| stateEncoding | \"text\"      | The state that goes out is always a string: a string is sent unchanged (not trimmed), any other value as its JSON text, and a state that is blank after trimming is refused before any request as an invalid_request. Omitted means the state goes out as given                              |\n\nThe rest of the entry is the same kind of data every decide entry carries: the\nsystem-one dialect, order 8, the endpoint {baseURL}/v1/scaledown, the model in\nthe body, trailing slashes dropped from the base URL, and the error rules in the\ntable below. The fields are also described in the generated API reference, under\nDecideBlock.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"How the catalog entry says this","lvl3":""}},
+{"objectID":"9f3c9d52f412bb381d8f770fa11e1de1c1f66ba5b56c324de62a11d5f6cf35bd","title":"Images and video","url":"/docs/getting-started/providers/scaledown#images-and-video","content":"ScaleDown's endpoint reads text only, and the catalog entry declares no media. An\nimage or a video attached to a decide() call for ScaleDown is refused before any\nrequest, as an invalid_request, and the catalog marks both models\nvision: false. This refusal comes before the blank-state check, so a call that\ncarries media is always refused for the media. To decide over images today, use\nXOR, Perplexity,\nCloudflare Clef or OpenAI Decisions.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"Images and video","lvl3":""}},
+{"objectID":"32661eaf2994628038e0d49fc81ee5e962c42273147ac024b6b6a66c0d2f2142","title":"Timeouts","url":"/docs/getting-started/providers/scaledown#timeouts","content":"The default is 5,000 ms per attempt (the shared decide default,\nDEFAULT_DECISION_TIMEOUT_MS). A choice question is scored once per option, and\nthe marketing session saw one ScaleDown 504 (Endpoint request timed out) after\nabout 29 seconds. ScaleDown's own latency is not measured here, so a caller with\nmany options or a long state should pass timeoutMs.\n\nA 504 that ScaleDown sends is classified timeout and is not retried. A timeout\nraised by NeuroLink's own clock is retried once.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"Timeouts","lvl3":""}},
+{"objectID":"a09b830ff06ece8be11a2d4fb2093b734e19a11ba1ba8393b17cce73b3f653b7","title":"Cost","url":"/docs/getting-started/providers/scaledown#cost","content":"List price: USD 0.05 per million input tokens, with output tokens not\ncharged, from ScaleDown's launch post and from the usage.cost in a response, as\nreported to this project by the marketing session on 2026-10-15. This project has\nnot priced a call itself and has not checked the rate since, so check ScaleDown's\ndashboard for the current one.\n\nNeuroLink prices a decision from the input tokens the reply reports, at that one\nrate, and counts output tokens at nothing. It does not read the reply's\nusage.cost. result.usage and the CLI's Cost: line are therefore an estimate\nfrom the list price, not a bill. A reply that reports no usage is priced at zero.\nThe sign-up allowance is a free allowance whose expiry is not established.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"Cost","lvl3":""}},
+{"objectID":"9fe742ca7befe0c2159befcd8359f9831dcb2b94f4f462faff3e74b86b47fffc","title":"Errors","url":"/docs/getting-started/providers/scaledown#errors","content":"ScaleDown's documentation lists HTTP 402 (insufficient credits), 422 (invalid),\n502 (model service error) and 504 (timeout). The marketing session also saw, on\n2026-10-15, a 429 Max queue length has been reached (no rate-limit headers; it\ncleared after a few seconds) and a 504 Endpoint request timed out, and each\nwas answered when it retried by hand. NeuroLink shows the detail text of an\nerror body, whether it is a plain string or an object with a message. A body in\nany other form, such as a detail that is an array, gives only\nScaleDown request failed with HTTP <status>; the kind and the retry behaviour\ndo not change.\n\n| Reply                                                | Kind                                                               | Retried   |\n| ---------------------------------------------------- | ------------------------------------------------------------------ | --------- |\n| 401                                                  | authentication — the provider instance stops retrying            | no        |\n| 402                                                  | invalid_request — the account has no credit                      | no        |\n| 422                                                  | invalid_request, with ScaleDown's reason when the body shows one | no        |\n| 429                                                  | rate_limit                                                       | yes, once |\n| 502                                                  | server                                                           | yes, once |\n| 504                                                  | timeout                                                          | no        |\n| other 5xx                                            | server                                                           | yes, once |\n| any other status                                     | invalid_request                                                  | no        |\n| no","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"Errors","lvl3":""}},
+{"objectID":"586e2eeaa1cb6c511d8eabf5f8e3aea68d19a0585eef1879b0a8432342bd93a7","title":"Data retention","url":"/docs/getting-started/providers/scaledown#data-retention","content":"NeuroLink has not read ScaleDown's data-handling terms. The state and the questions\nof every request go to api.scaledown.xyz, including the ones the built-in\nconsumers send for model routing, compaction, tool routing and RAG planning\nwhenever ScaleDown is the default decision provider. Read ScaleDown's terms before\nsending anything sensitive.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"Data retention","lvl3":""}},
+{"objectID":"4109a9812400e518479a4cdd59e777cb617fb3bff2364b69dae5e5fd397ecb04","title":"Not measured","url":"/docs/getting-started/providers/scaledown#not-measured","content":"Each item below is not measured by NeuroLink: nothing it was built or tested\nfrom says what the endpoint does beyond the marketing session's probes, and no\nrequest was made to ScaleDown for this project.\nContext size and input ceiling, for either model. The catalog records a\n   contextWindow of 128000 and a maxOutputTokens of 1 for classify-1, and\n   both are placeholders the catalog schema requires. 128000 is NeuroLink's own\n   default window for a model it knows nothing about, and 1 is the figure SubQ's\n   entry uses for a model that returns no text. Neither is a ScaleDown statement,\n   and the entry declares no decisionLimits, so nothing is refused locally on\n   the size of a state.\nLatency, for any size of state or number of options. The 5-second default\n   is the shared decide default, not a measurement of ScaleDown.\nRate limits. A 429 Max queue length has been reached, with no\n   rate-limit headers, was seen by the marketing session on 2026-10-15; no limit\n   is mapped.\nThe maximum number of options in a choice, of questions per request, and\n   of levels in a score, and what ScaleDown does with a list longer than the\n   2 to 10 levels its documentation describes.\nWhether classify moves to a newer model, and what that would change in an\n   answer.\nAccuracy of its answers against any other decide provider, and whether its\n   confidences are calibrated.\nHow long the free allowance lasts. Its expiry is not established.\nWhether a request ScaleDown refuses (an HTTP 502, 422 or 504) is charged.\nScaleDown's data-retention terms.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"Not measured","lvl3":""}},
+{"objectID":"ca6f431704d514c092d35d90635e8dc3bc58e2dc4c4e9d878edd8d0c9f95f466","title":"What has and has not been checked","url":"/docs/getting-started/providers/scaledown#what-has-and-has-not-been-checked","content":"Read or probed by the marketing session on 2026-10-15, not by this project:\n  ScaleDown's documentation (the endpoint, the model names, the question kinds,\n  the documented error statuses and the price); a request that sent the key as a\n  Bearer token answered HTTP 401 {\"detail\":\"x-api-key header is required\"}; a\n  model name other than classify or classify-1 answered HTTP 422; a state with\n  no text or document answered HTTP 422; the 255-option and 300-option choice\n  requests above; the 429 and the 504 above.\nIn this project: the whole behaviour of the entry is tested against stand-in\n  replies: the mocked provider suite (the request headers, body and URL, every\n  status above, the state rules, the order), the decide provider contract\n  (test/fixtures/decide/scaledown/contract.json, whose rows are all stand-ins\n  and none claims to have been seen live), and a CLI run against a local\n  server on 127.0.0.1. None of them reaches ScaleDown.\nNot captured: any real reply from ScaleDown. The catalog entry records no\n  live matrix.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"What has and has not been checked","lvl3":""}},
+{"objectID":"3ea0ed3ccd27e74b56a4ebd0bf35dc7eeceb527711f8d862c00af089437485d9","title":"Troubleshooting","url":"/docs/getting-started/providers/scaledown#troubleshooting","content":"ScaleDown requires an API key — set SCALEDOWN_API_KEY, or pass\n  credentials.scaledown.apiKey. A blank value counts as unset.\nA 401 — ScaleDown rejected the key, and the message carries its text. Check\n  the key. neurolink.decide() builds a fresh provider for every call, so the\n  next call asks again; only a provider instance you keep stops sending after a\n  rejection.\nA 402 — ScaleDown's documentation says this means no credit is left. Add\n  credit and call again; the same instance works again.\nA 422 — ScaleDown refused the request. The message carries its reason when the\n  body shows one; a model name other than classify or classify-1 is one cause.\nScaleDown needs non-blank state text — the state was an empty string or\n  only whitespace. Nothing was sent.\nScaleDown does not accept images or video — ScaleDown reads text only. Use\n  a decision provider that reads images.\nA 502 on a choice question with very many options — a probe by the\n  marketing session saw a 300-option choice refused this way. Send fewer options.\nA 504 / timeout — see Timeouts. Pass a larger timeoutMs.\nA request to /v1/v1/scaledown — SCALEDOWN_BASE_URL ends in /v1. Drop it.\nBuilt-in routing never uses ScaleDown — an earlier provider is also\n  configured and takes precedence (see When NeuroLink uses it),\n  or no key is set.\nA question is missing from result.answers — the reply left it out, or its\n  answer could not be read. An answer that could not be read is logged as a\n  warning that names the question and never quotes the reply; a question the\n  reply leaves out entirely is not logged, so look for it in result.answers.","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"Troubleshooting","lvl3":""}},
+{"objectID":"00d81ba62df89546669e56393ea60060a6cda460a8332e4f932f98c4126a49cc","title":"See also","url":"/docs/getting-started/providers/scaledown#see-also","content":"The decide inference type\nTypeSafe (Jev) Provider Guide\nLaya Provider Guide\nXOR Provider Guide\nSubQ Provider Guide\nPerplexity Decisions Provider Guide\nCloudflare Clef Provider Guide\nOpenAI Decisions Provider Guide\nScaleDown's documentation","hierarchy":{"lvl0":"Getting Started","lvl1":"ScaleDown Provider Guide","lvl2":"See also","lvl3":""}},
 {"objectID":"ba0463bfd09a65fed9ca1abc4022a8e8e54cbe54859a04dd2d4ae6b46eb7118e","title":"Scaleway Provider Guide","url":"/docs/getting-started/providers/scaleway","content":"Scaleway Provider Guide\n\nScaleway is a Tier-2 catalog provider: an OpenAI-compatible endpoint\ndescribed by one JSON file (src/lib/providers/catalog/scaleway.json) rather\nthan hand-written code. That file is the source of truth for everything on this\npage.\n\nVerification status: this entry is docs-verified only, not yet\nlive-verified. Scaleway's GET /v1/models needs a key (an unauthenticated\nGET answers HTTP 401), so the model ids come from the vendor's public\nsupported-models page and the roster has not been checked against the API. No\naccount was created and no API key was used to build this entry.\nevidence.liveMatrix is null until someone runs the live capability matrix\nwith a real key (see Verification status below).\n\nKey Facts\nProvider id: scaleway\nProtocol: OpenAI-compatible (/chat/completions)\nBase URL: https://api.scaleway.ai/v1 (the\n  Using Generative APIs page\n  also shows a project-scoped form, https://api.scaleway.ai/<project_id>/v1)\nDefault model: mistral-small-3.2-24b-instruct-2506\nModels in catalog: 11 — the chat models whose \"Available in Serverless?\"\n  cell on the\n  supported-models page\n  is Yes and that are not in that page's End of Life table. Left out: models\n  whose cell is No or EOL, the audio transcription and embedding models, and\n  six Serverless-Yes models that the End of Life table lists\n  (qwen3-coder-30b-a3b-instruct, pixtral-12b-2409, gemma-3-27b-it,\n  devstral-2-123b-instruct-2512, holo2-30b-a3b, voxtral-small-24b-2507)\nStreaming: supported — the\n  Chat Completions reference\n  documents stream with server-sent events and says the last chunk contains\n  data: [DONE]\nTool calling: model-dependent — the\n  function-calling page\n  says \"All the chat models hosted by Scaleway support function calling.\", the\n  supported-models page's attribute tables carry a \"Supports function calling\"\n  row that reads Yes for each of the eleven catalog ids, and for\n  gpt-oss-120b it says \"Currently, this model should be used through the\n  Responses API, because the Chat Completions API does not yet support\n  tool-calling for this model.\"\nTools while streaming: supported — the function-calling page has a \"Tool\n  calling with stream mode\" section\nStructured output: supported — the\n  structured-outputs page\n  documents json_schema (schema mode) and json_object (JSON mode) and says\n  \"All LLMs in the Scaleway library support Structured outputs and JSON mode.\"\nStructured output + tools together: not declared (false) — no combined\n  probe was possible without credentials\nEmbeddings: not declared. Scaleway's compatibility page documents an\n  Embeddings API and the supported-models page lists embedding models; this\n  entry covers chat models\nThinking: not declared (false) — the supported-models page lists\n  \"Supported reasoning efforts\" for eight of the catalog models (see the table\n  below); the provider-level flag stays false\nBilling: free-with-card (NeuroLink's classification, not a Scaleway\n  statement) — see Billing below\nKey format: none declared. The vendor's examples call the value\n  SCW_SECRET_KEY\n\nQuick Start\nGet an API key\nVisit: https://account.scaleway.com/register?service=console and create a Scaleway account; the quickstart (https://www.scaleway.com/en/docs/generative-apis/quickstart/) lists \"A Scaleway account logged into the console\" and \"A valid API key\" as requirements\nCreate an API key (https://www.scaleway.com/en/docs/iam/credentials/create-api-keys/): click IAM & API keys on the top-right drop-down menu of the Scaleway console, open the API keys tab and click + Generate API key. The page says: \"A screen displays showing the access key and secret key for your new API key and reminding you that this is your only chance to securely save the secret key.\" The quickstart passes the secret key as the OpenAI client's api_key. Querying models requires one of the permission sets GenerativeApisModelAccess, GenerativeApisFullAccess or AllProductsFullAccess (https://www.scaleway.com/en/docs/generative-apis/api-cli/using-generative-apis/)\nBilling as the vendor states it: the pricing page (https://www.scaleway.com/en/pricing/model-as-a-service/) says \"Try out new models with our free tier: 1 million tokens and 60 minutes of audio transcription.\"; the FAQ (https://www.scaleway.com/en/docs/generative-apis/faq/) says \"After reaching this limit, you will be charged per million tokens processed and per minutes of audio processed.\"; the create-account page (https://www.scaleway.com/en/docs/account/how-to/create-an-account/) says \"Ordering Scaleway resources requires a valid credit card.\"; the payment-method page (https://www.scaleway.com/en/docs/billing/how-to/add-payment-method/) says \"Before you can order Scaleway resources, you must add a payment method to your account.\"; the rate-limits page (https://www.scaleway.com/en/docs/generative-apis/reference-content/rate-limits/) says \"Base limits apply if you registered a valid payment method, and they are increased automatically if you also verify your identit","hierarchy":{"lvl0":"Getting Started","lvl1":"Scaleway Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"080be05f3ab9716f18b156ceb2ac1cbc0522a29a2479d922c0e6c0033e977231","title":"Scaleway Provider Guide","url":"/docs/getting-started/providers/scaleway#scaleway-provider-guide","content":"Scaleway is a Tier-2 catalog provider: an OpenAI-compatible endpoint\ndescribed by one JSON file (src/lib/providers/catalog/scaleway.json) rather\nthan hand-written code. That file is the source of truth for everything on this\npage.\n\nVerification status: this entry is docs-verified only, not yet\nlive-verified. Scaleway's GET /v1/models needs a key (an unauthenticated\nGET answers HTTP 401), so the model ids come from the vendor's public\nsupported-models page and the roster has not been checked against the API. No\naccount was created and no API key was used to build this entry.\nevidence.liveMatrix is null until someone runs the live capability matrix\nwith a real key (see Verification status below).","hierarchy":{"lvl0":"Getting Started","lvl1":"Scaleway Provider Guide","lvl2":"Scaleway Provider Guide","lvl3":""}},
 {"objectID":"cbfd9bee3a147c4e850d6008327e9708c8b1f6573a9a1de4069569df2638dc5e","title":"Key Facts","url":"/docs/getting-started/providers/scaleway#key-facts","content":"Provider id: scaleway\nProtocol: OpenAI-compatible (/chat/completions)\nBase URL: https://api.scaleway.ai/v1 (the\n  Using Generative APIs page\n  also shows a project-scoped form, https://api.scaleway.ai/<project_id>/v1)\nDefault model: mistral-small-3.2-24b-instruct-2506\nModels in catalog: 11 — the chat models whose \"Available in Serverless?\"\n  cell on the\n  supported-models page\n  is Yes and that are not in that page's End of Life table. Left out: models\n  whose cell is No or EOL, the audio transcription and embedding models, and\n  six Serverless-Yes models that the End of Life table lists\n  (qwen3-coder-30b-a3b-instruct, pixtral-12b-2409, gemma-3-27b-it,\n  devstral-2-123b-instruct-2512, holo2-30b-a3b, voxtral-small-24b-2507)\nStreaming: supported — the\n  Chat Completions reference\n  documents stream with server-sent events and says the last chunk contains\n  data: [DONE]\nTool calling: model-dependent — the\n  function-calling page\n  says \"All the chat models hosted by Scaleway support function calling.\", the\n  supported-models page's attribute tables carry a \"Supports function calling\"\n  row that reads Yes for each of the eleven catalog ids, and for\n  gpt-oss-120b it says \"Currently, this model should be used through the\n  Responses API, because the Chat Completions API does not yet support\n  tool-calling for this model.\"\nTools while streaming: supported — the function-calling page has a \"Tool\n  calling with stream mode\" section\nStructured output: supported — the\n  structured-outputs page\n  documents json_schema (schema mode) and json_object (JSON mode) and says\n  \"All LLMs in the Scaleway library support Structured outputs and JSON mode.\"\nStructured output + tools together: not declared (false) — no combined\n  probe was possible without credentials\nEmbeddings: not declared.","hierarchy":{"lvl0":"Getting Started","lvl1":"Scaleway Provider Guide","lvl2":"Key Facts","lvl3":""}},
@@ -7261,20 +7289,20 @@
 {"objectID":"96e991d6ea0464b1707e71a219f12f335400ad2b5401eaa86b74abb223adb7b3","title":"Verification status","url":"/docs/getting-started/providers/subconscious#verification-status","content":"Tier-2 onboarding requires evidence before a provider is accepted, and\npnpm run verify:provider-onboarding gates it in CI. This is what the\ncatalog currently records for Subconscious — docs- and roster-verified, not\nlive-verified:\n\n| Probe                     | Result                                                                                                                                                                                                                                                                                                                                               |\n| ------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Roster                    | unauthenticated GET https://api.subconscious.dev/v1/models, HTTP 200, 4 models, 2026-09-29 — no key used                                                                                                                                                                                                                                           |\n| Route check               | unauthenticated GET https://api.subconscious.dev/v1/chat/completions answered HTTP 405 with allow: POST, 2026-09-29; no POST was sent                                                                                                                                                                                                          |\n| Billing                   | public pages read without an account: the cloud API page and the pricing page.","hierarchy":{"lvl0":"Getting Started","lvl1":"Subconscious Provider Guide","lvl2":"Verification status","lvl3":""}},
 {"objectID":"8c3a0849c95eedb67f4cc7f557d8e1be167e442bfbc369468d98442b1312777e","title":"Error statuses","url":"/docs/getting-started/providers/subconscious#error-statuses","content":"The OpenAPI spec (https://docs.subconscious.dev/api-reference/openapi.json)\nlists these responses for /chat/completions:\n\n| Status | Vendor description                                           | NeuroLink class   |\n| ------ | ------------------------------------------------------------ | ----------------- |\n| 400    | Bad request due to invalid parameters or unsupported options | not in errorRules |\n| 401    | Unauthorized because of a missing or invalid API key         | authentication  |\n| 429    | Rate limited due to too many requests or tokens per minute   | rate-limit      |\n| 500    | Internal server error                                        | not in errorRules |","hierarchy":{"lvl0":"Getting Started","lvl1":"Subconscious Provider Guide","lvl2":"Error statuses","lvl3":""}},
 {"objectID":"0fca22eccff731488298cebb313d9ddfc374ef3c7fccdc469899e5119d9f1ad3","title":"See also","url":"/docs/getting-started/providers/subconscious#see-also","content":"Provider setup overview\nAll providers\nTier-2 onboarding — how this provider's JSON becomes a working integration\nProvider feature compatibility","hierarchy":{"lvl0":"Getting Started","lvl1":"Subconscious Provider Guide","lvl2":"See also","lvl3":""}},
-{"objectID":"7afe6dc192c3cc3ce8d4789877a559228c35059789628947b0c558d640232775","title":"SubQ Provider Guide","url":"/docs/getting-started/providers/subq","content":"SubQ Provider Guide\n\nA provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR,\nPerplexity and Cloudflare Clef,\nfrom Subquadratic's hosted SubQ Snap. It emits no text at all.\n\nOverview\n\nsubq-snap-preview is Subquadratic's decision model, in public preview. You\nsend one state plus named, typed questions. The model answers every question\nin a single pass and returns a typed answer for each, with a probability\ninstead of a sentence. SubQ speaks the same System One wire as TypeSafe, so\nNeuroLink describes it with one catalog file,\nsrc/lib/providers/catalog/subq.json, and no provider code of its own. NeuroLink\ncalls SubQ's public endpoint, so a key alone configures it: there is no base\nURL to set.\n\nKey Facts\n\n|                       |                                                                                                                                                                                    |\n| --------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Provider id           | subq; the aliases subquadratic and subq-snap name it too                                                                                                                     |\n| Inference type        | decide only                                                                                                                                                                      |\n| Model                 | subq-snap-preview, the only model on SubQ's Decisions endpoint. SUBQ_MODEL sets it                                                                                             |\n| Key                   | SUBQ_API_KEY, or credentials.subq.apiKey. Keys come from console.subq.ai; the console shows one only once                                           |\n| Endpoint              | https://api.subq.ai/v1/systemone; SUBQ_BASE_URL or credentials.subq.baseURL can name another origin                                                                          |\n| Questions per request | no cap declared                                                                                                                                                                    |\n| State window          | 376,000 estimated tokens, advisory: reported by decisionLimits(), never enforced. SubQ's own 400 response states a maximum model length of 393,216 tokens; see Limits |\n| Media                 | none. Images attached to a decide() call are refused before any request; see Images                                                                                   |\n| Default timeout       | 5 seconds; timeoutMs or --timeout replaces it                                                                                                                                  |\n| Precedence            | used automatically unless TypeSafe, Laya or XOR is configured; it comes ahead of Perplexity and Cloudflare Clef                                                                    |\n\nQuick Start\nGet a key\n\nSign up at console.subq.ai and create an API key. The\nconsole shows it once, so copy it then. SubQ answers HTTP 402 to a request made\nwith no credits left.\nConfigure\n\nSet the key in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:\n\nSUBQ_BASE_URL and credentials.subq.baseURL name an origin. NeuroLink appends\n/v1/systemone and drops any trailing slashes, so a value that already ends in\n/v1 would send the request to /v1/v1/systemone.\nUse it\n\nFrom the CLI\n\nWhen NeuroLink uses it\n\nEvery built-in consumer of decide — model routing, relevance-driven\ncompaction, tool routing and RAG planning — asks for the default decision\nprovider. That is the first one that is configured, in the environment or in\nthe credentials passed to the SDK, in the order TypeSafe, Laya,\nXOR, SubQ, Perplexity, then\nCloudflare Clef. SubQ counts with its key alone. It sits\nahead of Perplexity and Clef because its key belongs to no other provider, so\nsetting it can only be meant for SubQ, whereas Perplexity's and Clef's are shared\nwith text providers and must not displace it. So:\nOnly a SubQ key: built-in features use SubQ.\nA SubQ key, plus a Perplexity key or the Cloudflare pair: built-in\n  features use SubQ; Perplexity and Cloudflare Clef run only where a caller asks\n  for provider: \"perplexity-decider\" or provider: \"cloudflare-clef\".\nA SubQ key, plus TypeSafe's key, or Laya's or XOR's key and base URL:\n  built-in features use TypeSafe, Laya or XOR, in that order, whichever is\n  configured; SubQ runs only where a caller asks for provider: \"subq\".\nA blank SUBQ_API_KEY: no key. SubQ is not configured.\n\nLimits\n\nThe state window is advisory. decisionLimits() reports 376,000 estimated\nstate tokens, and t","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"","lvl3":""}},
+{"objectID":"7afe6dc192c3cc3ce8d4789877a559228c35059789628947b0c558d640232775","title":"SubQ Provider Guide","url":"/docs/getting-started/providers/subq","content":"SubQ Provider Guide\n\nA provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR,\nPerplexity and Cloudflare Clef,\nfrom Subquadratic's hosted SubQ Snap. It emits no text at all.\n\nOverview\n\nsubq-snap-preview is Subquadratic's decision model, in public preview. You\nsend one state plus named, typed questions. The model answers every question\nin a single pass and returns a typed answer for each, with a probability\ninstead of a sentence. SubQ speaks the same System One wire as TypeSafe, so\nNeuroLink describes it with one catalog file,\nsrc/lib/providers/catalog/subq.json, and no provider code of its own. NeuroLink\ncalls SubQ's public endpoint, so a key alone configures it: there is no base\nURL to set.\n\nKey Facts\n\n|                       |                                                                                                                                                                                    |\n| --------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Provider id           | subq; the aliases subquadratic and subq-snap name it too                                                                                                                     |\n| Inference type        | decide only                                                                                                                                                                      |\n| Model                 | subq-snap-preview, the only model on SubQ's Decisions endpoint. SUBQ_MODEL sets it                                                                                             |\n| Key                   | SUBQ_API_KEY, or credentials.subq.apiKey. Keys come from console.subq.ai; the console shows one only once                                           |\n| Endpoint              | https://api.subq.ai/v1/systemone; SUBQ_BASE_URL or credentials.subq.baseURL can name another origin                                                                          |\n| Questions per request | no cap declared                                                                                                                                                                    |\n| State window          | 376,000 estimated tokens, advisory: reported by decisionLimits(), never enforced. SubQ's own 400 response states a maximum model length of 393,216 tokens; see Limits |\n| Media                 | none. Images attached to a decide() call are refused before any request; see Images                                                                                   |\n| Default timeout       | 5 seconds; timeoutMs or --timeout replaces it                                                                                                                                  |\n| Precedence            | used automatically unless TypeSafe, Laya or XOR is configured; it comes ahead of Perplexity, Cloudflare Clef, OpenAI Decisions and ScaleDown                                       |\n\nQuick Start\nGet a key\n\nSign up at console.subq.ai and create an API key. The\nconsole shows it once, so copy it then. SubQ answers HTTP 402 to a request made\nwith no credits left.\nConfigure\n\nSet the key in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:\n\nSUBQ_BASE_URL and credentials.subq.baseURL name an origin. NeuroLink appends\n/v1/systemone and drops any trailing slashes, so a value that already ends in\n/v1 would send the request to /v1/v1/systemone.\nUse it\n\nFrom the CLI\n\nWhen NeuroLink uses it\n\nEvery built-in consumer of decide — model routing, relevance-driven\ncompaction, tool routing and RAG planning — asks for the default decision\nprovider. That is the first one that is configured, in the environment or in\nthe credentials passed to the SDK, in the order TypeSafe, Laya,\nXOR, SubQ, Perplexity,\nCloudflare Clef, OpenAI Decisions,\nthen ScaleDown. SubQ counts with its key alone. It sits\nahead of Perplexity, Clef and OpenAI Decisions because its key belongs to no other\nprovider, so setting it can only be meant for SubQ, whereas theirs are shared\nwith text providers and must not displace it. ScaleDown, whose key is also its\nown, was placed last on purpose. So:\nOnly a SubQ key: built-in features use SubQ.\nA SubQ key, plus a Perplexity key, the Cloudflare pair, OPENAI_API_KEY or\n  SCALEDOWN_API_KEY: built-in features use SubQ; Perplexity, Cloudflare Clef,\n  OpenAI Decisions and ScaleDown run only where a caller asks for\n  provider: \"perplexity-decider\", provider: \"cloudflare-clef\",\n  provider: \"openai-decisions\" or provider: \"scaledown\".\nA SubQ key, plus TypeSafe's key, or Laya's or XOR's key and base URL:\n  built-in features use TypeSafe, Laya or XOR, in that order, whichever is\n  configu","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"e3bea9ca02b054acb28554849e0e81ca2881ff4789c51ecfae77359c4de17488","title":"SubQ Provider Guide","url":"/docs/getting-started/providers/subq#subq-provider-guide","content":"A provider of decide — the same typed boolean / choice / score\nanswers as TypeSafe's Jev, Laya, XOR,\nPerplexity and Cloudflare Clef,\nfrom Subquadratic's hosted SubQ Snap. It emits no text at all.","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"SubQ Provider Guide","lvl3":""}},
 {"objectID":"e9a1296ef2e945c77e18e49fd508f9925268bcba0dbc91fd332f0d303f3bed67","title":"Overview","url":"/docs/getting-started/providers/subq#overview","content":"subq-snap-preview is Subquadratic's decision model, in public preview. You\nsend one state plus named, typed questions. The model answers every question\nin a single pass and returns a typed answer for each, with a probability\ninstead of a sentence. SubQ speaks the same System One wire as TypeSafe, so\nNeuroLink describes it with one catalog file,\nsrc/lib/providers/catalog/subq.json, and no provider code of its own. NeuroLink\ncalls SubQ's public endpoint, so a key alone configures it: there is no base\nURL to set.","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"Overview","lvl3":""}},
 {"objectID":"1ba4213315360e1bd7ffb4d6a6155269f05ca6642867dc7f9345ed30ad0dcb87","title":"Key Facts","url":"/docs/getting-started/providers/subq#key-facts","content":"|                       |                                                                                                                                                                                    |\n| --------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Provider id           | subq; the aliases subquadratic and subq-snap name it too                                                                                                                     |\n| Inference type        | decide only                                                                                                                                                                      |\n| Model                 | subq-snap-preview, the only model on SubQ's Decisions endpoint. SUBQ_MODEL sets it                                                                                             |\n| Key                   | SUBQ_API_KEY, or credentials.subq.apiKey. Keys come from console.subq.ai; the console shows one only once                                           |\n| Endpoint              | https://api.subq.ai/v1/systemone; SUBQ_BASE_URL or credentials.subq.baseURL can name another origin                                                                          |\n| Questions per request | no cap declared                                                                                                                                                                    |\n| State window          | 376,000 estimated tokens, advisory: reported by decisionLimits(), never enforced. SubQ's own 400 response states a maximum model length of 393,216 tokens; see Limits |\n| Media                 | none.","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"Key Facts","lvl3":""}},
 {"objectID":"b71131b7794e8441aa8055e684d8b8585260f61cd71db50032c1006f2f32306a","title":"1. Get a key","url":"/docs/getting-started/providers/subq#1-get-a-key","content":"Sign up at console.subq.ai and create an API key. The\nconsole shows it once, so copy it then. SubQ answers HTTP 402 to a request made\nwith no credits left.","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"1. Get a key","lvl3":""}},
 {"objectID":"1701bc3f69f008aee8581f0a43107253347a3ae74646648b87336e8fdc63dd50","title":"2. Configure","url":"/docs/getting-started/providers/subq#2-configure","content":"Set the key in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:\n\nSUBQ_BASE_URL and credentials.subq.baseURL name an origin. NeuroLink appends\n/v1/systemone and drops any trailing slashes, so a value that already ends in\n/v1 would send the request to /v1/v1/systemone.","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"2. Configure","lvl3":""}},
-{"objectID":"3827eaed3d77845619f0aca3fcb31fa406bcb4130bf489035a4e673470170ffb","title":"When NeuroLink uses it","url":"/docs/getting-started/providers/subq#when-neurolink-uses-it","content":"Every built-in consumer of decide — model routing, relevance-driven\ncompaction, tool routing and RAG planning — asks for the default decision\nprovider. That is the first one that is configured, in the environment or in\nthe credentials passed to the SDK, in the order TypeSafe, Laya,\nXOR, SubQ, Perplexity, then\nCloudflare Clef. SubQ counts with its key alone. It sits\nahead of Perplexity and Clef because its key belongs to no other provider, so\nsetting it can only be meant for SubQ, whereas Perplexity's and Clef's are shared\nwith text providers and must not displace it. So:\nOnly a SubQ key: built-in features use SubQ.\nA SubQ key, plus a Perplexity key or the Cloudflare pair: built-in\n  features use SubQ; Perplexity and Cloudflare Clef run only where a caller asks\n  for provider: \"perplexity-decider\" or provider: \"cloudflare-clef\".\nA SubQ key, plus TypeSafe's key, or Laya's or XOR's key and base URL:\n  built-in features use TypeSafe, Laya or XOR, in that order, whichever is\n  configured; SubQ runs only where a caller asks for provider: \"subq\".\nA blank SUBQ_API_KEY: no key. SubQ is not configured.","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"When NeuroLink uses it","lvl3":""}},
+{"objectID":"3827eaed3d77845619f0aca3fcb31fa406bcb4130bf489035a4e673470170ffb","title":"When NeuroLink uses it","url":"/docs/getting-started/providers/subq#when-neurolink-uses-it","content":"Every built-in consumer of decide — model routing, relevance-driven\ncompaction, tool routing and RAG planning — asks for the default decision\nprovider. That is the first one that is configured, in the environment or in\nthe credentials passed to the SDK, in the order TypeSafe, Laya,\nXOR, SubQ, Perplexity,\nCloudflare Clef, OpenAI Decisions,\nthen ScaleDown. SubQ counts with its key alone. It sits\nahead of Perplexity, Clef and OpenAI Decisions because its key belongs to no other\nprovider, so setting it can only be meant for SubQ, whereas theirs are shared\nwith text providers and must not displace it. ScaleDown, whose key is also its\nown, was placed last on purpose. So:\nOnly a SubQ key: built-in features use SubQ.\nA SubQ key, plus a Perplexity key, the Cloudflare pair, OPENAI_API_KEY or\n  SCALEDOWN_API_KEY: built-in features use SubQ; Perplexity, Cloudflare Clef,\n  OpenAI Decisions and ScaleDown run only where a caller asks for\n  provider: \"perplexity-decider\", provider: \"cloudflare-clef\",\n  provider: \"openai-decisions\" or provider: \"scaledown\".\nA SubQ key, plus TypeSafe's key, or Laya's or XOR's key and base URL:\n  built-in features use TypeSafe, Laya or XOR, in that order, whichever is\n  configured; SubQ runs only where a caller asks for provider: \"subq\".\nA blank SUBQ_API_KEY: no key. SubQ is not configured.","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"When NeuroLink uses it","lvl3":""}},
 {"objectID":"1d299f5eff73ec92910a1c96e5e24965d2594c3f049ce8f0ba6a1d5dbb73bd02","title":"Limits","url":"/docs/getting-started/providers/subq#limits","content":"The state window is advisory. decisionLimits() reports 376,000 estimated\nstate tokens, and the catalog records a model contextWindow of 393,216. Both\ncome from one live observation made on 2026-10-13 for NeuroLink (not by this\nchange): SubQ refused a request of 399,795 tokens with HTTP 400, and the message\nnames a maximum model length of 393,216 tokens; the largest request it answered\nhad 376,180 input tokens. So 393,216 is the maximum model length SubQ's own 400\nresponse states, and 376,000 is a margin under the largest request it answered.\nBoth are advisory: a longer state is sent, the server decides, and neither\nfigure is a guarantee. The 376,000 counts NeuroLink's estimated tokens, which can\ndiffer from the tokens SubQ counts, so a state near it can still be refused.\n\nSubQ answers an over-length request with HTTP 400 and a validation_error body\n(The decoder prompt (length 399795) is longer than the maximum model length of\n393216 ...). NeuroLink reports it as invalid_request, not retried; it is not\nreported as max_tokens_exceeded.\n\nNo question cap is declared. SubQ's documentation gives no count; a choice\nquestion takes 2 to 255 options, a score question 2 to 10 levels, and a\nquestion name up to 100 characters. SubQ refuses a request outside those with\nHTTP 422, which NeuroLink reports as invalid_request, with SubQ's reason when\nthe body is the documented object form.\n\nTimeout. The default is 5 seconds per attempt. A state near the window takes\nlonger to read than a short one, so size timeoutMs for the states you send.\n\nPrice. SubQ's documentation gives $0.05 per million input tokens and no\ncharge for output tokens (read on 2026-10-13; check the console for the current\nrate). Each response carries usage.input_tokens, which decide() returns as\nusage.inputTokens.","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"Limits","lvl3":""}},
 {"objectID":"4c664b622160929b5860a9d062387141d6a2902dade93c7d7d8887e13ff8fc9d","title":"Images","url":"/docs/getting-started/providers/subq#images","content":"SubQ's documentation describes image input: up to 4 images per request, each an\nobject with a mediaType and a one-line base64 string, with no data:\nprefix. NeuroLink's decide() prepares images as data: URLs, and a catalog\nentry cannot yet describe the object form, so SubQ's entry declares no media.\nAn image attached to a decide() call for SubQ is refused before any request,\nas invalid_request, and the catalog marks the model vision: false. To\ndecide over images today, use XOR, Perplexity\nor Cloudflare Clef.","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"Images","lvl3":""}},
 {"objectID":"789f51fadc3b52d17f55a04b15240e2a8c4b722dae7a88131e99a18447fe3568","title":"Errors","url":"/docs/getting-started/providers/subq#errors","content":"SubQ's documentation gives an error body of {\"detail\": {\"error_type\": \"…\",\n\"message\": \"…\"}}. NeuroLink shows the message, or a detail that is a plain\nstring. A body in any other form, such as a detail that is an array, gives only\nSubQ request failed with HTTP <status>; the kind and the retry behaviour do not\nchange.\n\n| Reply | Kind                                                                                       | Retried |\n| ----- | ------------------------------------------------------------------------------------------ | ------- |\n| 401   | authentication — the provider instance stops retrying                                    | no      |\n| 402   | invalid_request — the account has no credits                                             | no      |\n| 403   | invalid_request — the key cannot use the model                                           | no      |\n| 400   | invalid_request, with SubQ's reason (for example a prompt past the maximum model length) | no      |\n| 422   | invalid_request, with SubQ's reason (documented body form)                               | no      |\n| 429   | rate_limit                                                                               | yes     |\n| 503   | overloaded                                                                               | yes     |\n| 5xx   | server                                                                                   | yes     |\n| other | invalid_request                                                                          | no      |\n| image | invalid_request, with no network call (see Images)                            | no      |\n\nOnly a 401 stops the instance. A 402 is not treated as a rejected key on\npurpose: after you add credit, the same instance works again, with no restart.\nAn error message never carries the API key; if SubQ echoes it, NeuroLink\nreplaces it before the message reaches a log or an error.","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"Errors","lvl3":""}},
 {"objectID":"e2bf63ecccdffc248bdf0bee10df038bac383d491a3f3ce10694a3ec2acc7fa3","title":"Data retention","url":"/docs/getting-started/providers/subq#data-retention","content":"Check SubQ's data retention page\nbefore sending anything sensitive. SubQ's documentation, as read on 2026-10-13\n(a reading made before this page was written, and not repeated for it), says\nthat prompts are kept indefinitely unless the organization sets otherwise, and\nthat Decisions are unaffected by zero data retention. Check the page for the\ncurrent terms.","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"Data retention","lvl3":""}},
 {"objectID":"9034af0b1ace52d3498d4d50dc9c313deb0155f7177e46b1eb954b7b4a1afd06","title":"What has and has not been checked","url":"/docs/getting-started/providers/subq#what-has-and-has-not-been-checked","content":"Read on 2026-10-13, before this page was written and not repeated for it:\n  SubQ's documentation (the endpoint, the error body shape, the limits on\n  questions and images, the price and the retention terms), and an\n  unauthenticated GET https://api.subq.ai/v1/models that listed\n  subq-snap-preview.\nOne decision was captured: a single request from SubQ's documentation,\n  answered HTTP 200. Its body is the contract fixture's success body,\n  test/fixtures/decide/subq/contract.json.\nObserved live on 2026-10-13, in a run made for NeuroLink and not repeated\n  for this page: the over-length 400 above (its body is in the fixture as a\n  seen-live row; the saved text of the message stops after \"number of text\n  token\") and the largest request answered, 376,180 input tokens.\nNot captured: any other error body. Every other error in that fixture is a\n  stand-in written from the shape the documentation describes, and is labelled\n  so. Whether a 422 ever carries an array detail is unknown.\nNot measured: a question cap, and what an image costs. The catalog entry\n  records no live matrix.","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"What has and has not been checked","lvl3":""}},
 {"objectID":"e5c18a988b3c46629a67eff25416973546e1e390bfbfc68193e59afcc913b13e","title":"Troubleshooting","url":"/docs/getting-started/providers/subq#troubleshooting","content":"SubQ requires an API key — set SUBQ_API_KEY, or pass\n  credentials.subq.apiKey. A blank value counts as unset.\nAuthentication failed — SubQ answered 401. Check the key in the console,\n  then build a new NeuroLink instance: an instance does not retry after a\n  rejection.\nA 402 — the account has no credits. The message says so; add credit and\n  call again.\nA 403 — the key cannot use the model. The message says why.\nA request to /v1/v1/systemone — SUBQ_BASE_URL ends in /v1. It should\n  be the origin, https://api.subq.ai.\nBuilt-in routing never uses SubQ — another decision provider is also\n  configured and takes precedence.","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"Troubleshooting","lvl3":""}},
-{"objectID":"ef33539c4045e9eb9773ff73720a9ab965d7c4f0e27877d2ebb90670c4b63477","title":"See also","url":"/docs/getting-started/providers/subq#see-also","content":"The decide inference type\nTypeSafe (Jev) Provider Guide\nLaya Provider Guide\nXOR Provider Guide\nPerplexity Decisions Provider Guide\nCloudflare Clef Provider Guide\nSubQ's Decisions API","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"See also","lvl3":""}},
+{"objectID":"ef33539c4045e9eb9773ff73720a9ab965d7c4f0e27877d2ebb90670c4b63477","title":"See also","url":"/docs/getting-started/providers/subq#see-also","content":"The decide inference type\nTypeSafe (Jev) Provider Guide\nLaya Provider Guide\nXOR Provider Guide\nPerplexity Decisions Provider Guide\nCloudflare Clef Provider Guide\nOpenAI Decisions Provider Guide\nScaleDown Provider Guide\nSubQ's Decisions API","hierarchy":{"lvl0":"Getting Started","lvl1":"SubQ Provider Guide","lvl2":"See also","lvl3":""}},
 {"objectID":"5d6fb26e3e893c2534b09e38ae652750f963771afe087027bb853555b13eb8a4","title":"Synthetic Provider Guide","url":"/docs/getting-started/providers/synthetic","content":"Synthetic Provider Guide\n\nSynthetic is a Tier-2 catalog provider: OpenAI-wire-compatible with no\nbehavioural quirks declared, so its entire integration is one JSON file\n(src/lib/providers/catalog/synthetic.json) rather than hand-written code.\nThat file is the source of truth for everything on this page.\n\nVerification status: this entry is docs- and roster-verified, not yet\nlive-verified. Every model id, context window, output ceiling and price\nbelow comes from an unauthenticated GET /openai/v1/models call plus the\nvendor's own pricing page and LiteLLM's third-party provider page — no API\nkey was used or is required to onboard this entry. Nobody has yet run a\nlive capability sweep (chat, streaming, tool calls, structured output)\nagainst Synthetic with a real key; evidence.liveMatrix is null until\nthat happens. Treat capabilities.tools, .structuredOutput and\n.thinking as \"documented by the vendor,\" not \"wire-proven.\"\n\nKey Facts\nProvider id: synthetic\nProtocol: OpenAI-compatible (/chat/completions — the only operation\n  LiteLLM's provider page lists for Synthetic)\nBase URL: https://api.synthetic.new/openai/v1\nDefault model: syn:large:text (deepseek-ai/DeepSeek-V4.1-Flash)\nModels in catalog: 11 ids covering 7 unique always-on models (short\n  syn:* aliases and full Hugging Face weight ids both resolve)\nStreaming: documented (vendor docs show a streaming SDK example)\nTool calling: documented — every model on the live roster lists tools\n  in supported_features\nStructured output: documented — every model lists json_mode and\n  structured_outputs in supported_features. Tools + structured output\n  together has not been probed, so structuredOutputWithTools is false\n  until a live combined request confirms it.\nReasoning / thinking: documented — every model lists reasoning in\n  supported_features with an explicit reasoning_parameters.efforts list\n  (e.g. none/low/high/xhigh/max)\nEmbeddings: not exposed through this catalog entry. Synthetic's\n  pricing page separately lists an included embedding model\n  (hf:nomic-ai/nomic-embed-text-v1.5), but NeuroLink's embed() /\n  embedMany() surface is currently wired only for OpenAI, Google AI\n  Studio, Google Vertex and Amazon Bedrock — not generic Tier-2 catalog\n  providers — so capabilities.embeddings is false here regardless.\nBilling: no free tier — flat-fee subscription ($1/day or $30/month,\n  500 requests/5hr) or enterprise usage-based pricing\nKey format: none documented\n\nQuick Start\nGet an API key\nVisit: https://synthetic.new and choose Sign up (or Log in if you already have an account)\nSubscribe to a Pack ($1/day or $30/month, 500 requests/5hr per the pricing page) or set up enterprise usage-based billing — Synthetic has no free tier or keyless trial\nGenerate an API key from your account dashboard\nSet SYNTHETIC_API_KEY in your .env file\nConfigure\nUse it\n\nPer-request credentials work as they do for every provider:\n\nModels\n\n| Model                                               | Context | Max out | Vision | $/M in · out · cached-in | Notes                                                                                                           |\n| --------------------------------------------------- | ------- | ------- | ------ | ------------------------ | --------------------------------------------------------------------------------------------------------------- |\n| syn:large:text ⭐                                 | 512K    | 64K     | yes    | $0.60 / $1.20 / $0.03    | Recommended — alias for deepseek-ai/DeepSeek-V4.1-Flash. Vendor: \"an efficient near-Fable model with vision.\" |\n| syn:small:text                                    | 192K    | 64K     | no     | $0.10 / $0.50 / $0.02    | Alias for zai-org/GLM-4.7-Flash — smaller/faster, ~30B class; text-only input.                                |\n| syn:large:vision                                  | 512K    | 64K     | yes    | $3.00 / $15.00 / $0.45   | Alias for moonshotai/Kimi-K3 — 2.8T-parameter flagship; the most expensive model in the roster.               |\n| syn:small:vision                                  | 256K    | 64K     | yes    | $0.45 / $2.20 / $0.09    | Alias for Qwen/Qwen3.8-27B — small coding + vision model.                                                     |\n| hf:zai-org/GLM-5.3-Flash                          | 512K    | 64K     | yes    | $0.15 / $0.50 / $0.04    | No syn: short alias. Vendor: \"an extremely efficient, Opus-4.8-equivalent model with vision.\"                 |\n| hf:openai/gpt-oss-120b                            | 128K    | 64K     | no     | $0.10 / $0.10 / $0.02    | OpenAI's open-weight gpt-oss-120b; strong reasoning + agentic use.                                              |\n| hf:nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4 | 256K    | 64K     | no     | $0.30 / $1.00 / $0.06    | Fast 120B model, good for coding agents.                                                                        |\n| hf:deepseek-ai/DeepSeek-V4.1-Flash                | 512K    | 64K     | yes    | $0","hierarchy":{"lvl0":"Getting Started","lvl1":"Synthetic Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"eef94ef22cf845eb2e5117a994e6483af7f0a9b385fffa0c0b1155e6a836c314","title":"Synthetic Provider Guide","url":"/docs/getting-started/providers/synthetic#synthetic-provider-guide","content":"Synthetic is a Tier-2 catalog provider: OpenAI-wire-compatible with no\nbehavioural quirks declared, so its entire integration is one JSON file\n(src/lib/providers/catalog/synthetic.json) rather than hand-written code.\nThat file is the source of truth for everything on this page.\n\nVerification status: this entry is docs- and roster-verified, not yet\nlive-verified. Every model id, context window, output ceiling and price\nbelow comes from an unauthenticated GET /openai/v1/models call plus the\nvendor's own pricing page and LiteLLM's third-party provider page — no API\nkey was used or is required to onboard this entry. Nobody has yet run a\nlive capability sweep (chat, streaming, tool calls, structured output)\nagainst Synthetic with a real key; evidence.liveMatrix is null until\nthat happens. Treat capabilities.tools, .structuredOutput and\n.thinking as \"documented by the vendor,\" not \"wire-proven.\"","hierarchy":{"lvl0":"Getting Started","lvl1":"Synthetic Provider Guide","lvl2":"Synthetic Provider Guide","lvl3":""}},
 {"objectID":"c2baa496ddd36ed9f0523b9d3466fa858e3fe791889c16953111aa282b9ccbb5","title":"Key Facts","url":"/docs/getting-started/providers/synthetic#key-facts","content":"Provider id: synthetic\nProtocol: OpenAI-compatible (/chat/completions — the only operation\n  LiteLLM's provider page lists for Synthetic)\nBase URL: https://api.synthetic.new/openai/v1\nDefault model: syn:large:text (deepseek-ai/DeepSeek-V4.1-Flash)\nModels in catalog: 11 ids covering 7 unique always-on models (short\n  syn:* aliases and full Hugging Face weight ids both resolve)\nStreaming: documented (vendor docs show a streaming SDK example)\nTool calling: documented — every model on the live roster lists tools\n  in supported_features\nStructured output: documented — every model lists json_mode and\n  structured_outputs in supported_features. Tools + structured output\n  together has not been probed, so structuredOutputWithTools is false\n  until a live combined request confirms it.\nReasoning / thinking: documented — every model lists reasoning in\n  supported_features with an explicit reasoning_parameters.efforts list\n  (e.g. none/low/high/xhigh/max)\nEmbeddings: not exposed through this catalog entry. Synthetic's\n  pricing page separately lists an included embedding model\n  (hf:nomic-ai/nomic-embed-text-v1.5), but NeuroLink's embed() /\n  embedMany() surface is currently wired only for OpenAI, Google AI\n  Studio, Google Vertex and Amazon Bedrock — not generic Tier-2 catalog\n  providers — so capabilities.embeddings is false here regardless.\nBilling: no free tier — flat-fee subscription ($1/day or $30/month,\n  500 requests/5hr) or enterprise usage-based pricing\nKey format: none documented","hierarchy":{"lvl0":"Getting Started","lvl1":"Synthetic Provider Guide","lvl2":"Key Facts","lvl3":""}},
@@ -7331,19 +7359,19 @@
 {"objectID":"14c8755cfa45311a2b8645d3729d760f513b4e768f4cbe6d75e1dcaaa68c7e94","title":"Configuration Reference","url":"/docs/getting-started/providers/together-ai#configuration-reference","content":"| Environment Variable | Required | Default                                   |\n| -------------------- | -------- | ----------------------------------------- |\n| TOGETHER_API_KEY   | Yes      | —                                         |\n| TOGETHER_MODEL     | No       | meta-llama/Llama-3.3-70B-Instruct-Turbo |\n| TOGETHER_BASE_URL  | No       | https://api.together.xyz/v1             |","hierarchy":{"lvl0":"Getting Started","lvl1":"Together AI Provider Guide","lvl2":"Configuration Reference","lvl3":""}},
 {"objectID":"73dacb114377509534778a0d5d1d386e64fe4e71f75f590903d6a5c3dda075c6","title":"Feature Support Matrix","url":"/docs/getting-started/providers/together-ai#feature-support-matrix","content":"| Feature           | Support           |\n| ----------------- | ----------------- |\n| Text generation   | Yes               |\n| Streaming         | Yes               |\n| Tool calling      | Yes (model-dep.)  |\n| Structured output | Yes (model-dep.)  |\n| Vision            | Yes (Llama 3.2 V) |\n| Embeddings        | Limited           |","hierarchy":{"lvl0":"Getting Started","lvl1":"Together AI Provider Guide","lvl2":"Feature Support Matrix","lvl3":""}},
 {"objectID":"5a31d744d2d35af19cc87e27118e0d4057ab98d14fa7be6e9acb01dcbfc29f6a","title":"See Also","url":"/docs/getting-started/providers/together-ai#see-also","content":"Fireworks Provider\nGroq Provider","hierarchy":{"lvl0":"Getting Started","lvl1":"Together AI Provider Guide","lvl2":"See Also","lvl3":""}},
-{"objectID":"63e8301588e0fd846272e6b041d27fbe6a754543f38d949f07ce7e6f181a0ced","title":"TypeSafe (Jev) Provider Guide","url":"/docs/getting-started/providers/typesafe","content":"TypeSafe (Jev) Provider Guide\n\nOne of the providers that serve decide rather than generate/stream —\nit returns typed, calibrated judgments and emits no text at all. The others are\nLaya and XOR, open-weights models served from an endpoint\nyou configure, and SubQ, Perplexity,\nCloudflare Clef and OpenAI Decisions,\nhosted APIs. When more than one is configured, TypeSafe is tried before the\nothers.\n\nOverview\n\nTypeSafe's Jev is a \"System One\" model. You send one state plus a map of\nnamed, typed questions; it returns one typed answer per question, all evaluated\nin a single parallel pass. Nothing has to be parsed back out of prose, and every\nchoice/score answer carries a calibrated confidence rather than a\nself-reported one.\n\nBecause it emits no text, generate() and stream() are not available and\ngetAISDKModel() throws — the same shape Voyage and Jina already use for\nembedding-only providers. Its descriptor declares inferenceKinds: [\"decide\"],\nwhich keeps it out of auto-select and the health sweep, so those throws are\nunreachable in normal use.\n\nThis is not neurolink.evaluate(), which\nscores an already-generated response with RAGAS scorers. Different feature,\ndifferent word.\n\nKey Facts\nProvider id: typesafe (aliases: jev, typesafe-ai)\nInference kinds: decide only — one of the providers that do (the others\n  are Laya, XOR, SubQ,\n  Perplexity,\n  Cloudflare Clef and\n  OpenAI Decisions); TypeSafe is tried before them when\n  more than one is configured\nTool calling: none (toolSupport: \"none\") — a decision model calls nothing\nHealth check: env-only; it is never probed with a live generation\nDefault decide timeout: 5000 ms (timeouts.decideMs)\nLatency: flat in question count — 1 question ~393 ms, 400 questions\n  ~465 ms. Concurrent requests queue instead, so batch every question into one\n  call rather than fanning out.\nCost: ~$0.042 per million input tokens, output billed at zero — about\n  $0.00002 per decision. Output tokens are reported — measured 21 for a\n  single question, converging to ~17.5 per question in a batch of eight — they\n  are simply not charged.\nAccuracy is the trade: 67.8% on TypeSafe's own 711-case benchmark against\n  Opus 5's 73.1%. Right for decisions that are gated and reversible; wrong for\n  final answers.\n\nQuick Start\nGet an API key\n\nCreate one at console.typesafe.ai/keys.\nConfigure\nUse it\n\ntryDecide() returns null on any failure. Use decide() when you want the\nfailure to surface; it throws a ProviderError whose cause carries a typed\nkind.\n\nFrom the CLI\n\nneurolink decide [state] calls decide() under the hood and reads\nTYPESAFE_API_KEY from the environment like every other CLI command. See the\nCLI command reference.\n\nThe degradation contract\n\nSetting the key is the entire switch for TypeSafe, and removing it is a\ncomplete undo while no other decision provider is configured. Every internal\nconsumer of decide fails open: with no decision provider configured, model\nrouting, context budgeting, relevance compaction, tool routing and RAG planning\nall behave exactly as they did before. Note that Perplexity's key is shared with\nits text provider, so a PERPLEXITY_API_KEY set for that provider also counts as\na configured decision provider; see\nOne key, two providers. The same\ngoes for CLOUDFLARE_API_KEY with CLOUDFLARE_ACCOUNT_ID, which the Workers AI\ntext provider reads too: set together they configure Cloudflare Clef, which ranks\nafter TypeSafe, Laya, XOR, SubQ and Perplexity; see One token, two providers.\nOPENAI_API_KEY, which the OpenAI chat provider reads too, configures OpenAI\nDecisions, the last fallback; see\nOne key, two providers. There is no\nconfiguration in which a missing, invalid, slow or unreachable decision model\nchanges NeuroLink's observable behaviour.\n\nA credential the service does not accept disables that provider instance rather\nthan paying a round trip on every later call to be told so again.\n\nTwo transports\n\nThe same model is reachable two ways, and the choice is made once in the\nconstructor.\n\n|                     | Direct              | Vercel AI Gateway                             |\n| ------------------- | ------------------- | --------------------------------------------- |\n| Key                 | TYPESAFE_API_KEY  | AI_GATEWAY_API_KEY                          |\n| Endpoint            | api.typesafe.ai   | ai-gateway.vercel.sh/v4/ai/evaluation-model |\n| Endpoint override   | TYPESAFE_BASE_URL | TYPESAFE_GATEWAY_URL                        |\n| Model named in      | request body        | ai-model-id header                          |\n| Question vocabulary | noul              | boolean                                     |\n| confidence        | on each answer      | on providerMetadata                         |\n| Billed by           | TypeSafe            | Vercel                                        |\n\nHolding both keys keeps the direct transport, so the confidence figures a\nhost already sees do not shift underneath it when a second key appears. Force\none with TYPESAFE_TRANSPORT=direct|gat","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"","lvl3":""}},
-{"objectID":"7320dc5177f0e320fe7887f77c01c06519bb6c0ddeb8a8b38ecfa013a9363a1f","title":"TypeSafe (Jev) Provider Guide","url":"/docs/getting-started/providers/typesafe#typesafe-jev-provider-guide","content":"One of the providers that serve decide rather than generate/stream —\nit returns typed, calibrated judgments and emits no text at all. The others are\nLaya and XOR, open-weights models served from an endpoint\nyou configure, and SubQ, Perplexity,\nCloudflare Clef and OpenAI Decisions,\nhosted APIs. When more than one is configured, TypeSafe is tried before the\nothers.","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"TypeSafe (Jev) Provider Guide","lvl3":""}},
+{"objectID":"63e8301588e0fd846272e6b041d27fbe6a754543f38d949f07ce7e6f181a0ced","title":"TypeSafe (Jev) Provider Guide","url":"/docs/getting-started/providers/typesafe","content":"TypeSafe (Jev) Provider Guide\n\nOne of the providers that serve decide rather than generate/stream —\nit returns typed, calibrated judgments and emits no text at all. The others are\nLaya and XOR, open-weights models served from an endpoint\nyou configure, and SubQ, Perplexity,\nCloudflare Clef, OpenAI Decisions\nand ScaleDown, hosted APIs. When more than one is configured, TypeSafe is tried before the\nothers.\n\nOverview\n\nTypeSafe's Jev is a \"System One\" model. You send one state plus a map of\nnamed, typed questions; it returns one typed answer per question, all evaluated\nin a single parallel pass. Nothing has to be parsed back out of prose, and every\nchoice/score answer carries a calibrated confidence rather than a\nself-reported one.\n\nBecause it emits no text, generate() and stream() are not available and\ngetAISDKModel() throws — the same shape Voyage and Jina already use for\nembedding-only providers. Its descriptor declares inferenceKinds: [\"decide\"],\nwhich keeps it out of auto-select and the health sweep, so those throws are\nunreachable in normal use.\n\nThis is not neurolink.evaluate(), which\nscores an already-generated response with RAGAS scorers. Different feature,\ndifferent word.\n\nKey Facts\nProvider id: typesafe (aliases: jev, typesafe-ai)\nInference kinds: decide only — one of the providers that do (the others\n  are Laya, XOR, SubQ,\n  Perplexity,\n  Cloudflare Clef,\n  OpenAI Decisions and\n  ScaleDown); TypeSafe is tried before them when\n  more than one is configured\nTool calling: none (toolSupport: \"none\") — a decision model calls nothing\nHealth check: env-only; it is never probed with a live generation\nDefault decide timeout: 5000 ms (timeouts.decideMs)\nLatency: flat in question count — 1 question ~393 ms, 400 questions\n  ~465 ms. Concurrent requests queue instead, so batch every question into one\n  call rather than fanning out.\nCost: ~$0.042 per million input tokens, output billed at zero — about\n  $0.00002 per decision. Output tokens are reported — measured 21 for a\n  single question, converging to ~17.5 per question in a batch of eight — they\n  are simply not charged.\nAccuracy is the trade: 67.8% on TypeSafe's own 711-case benchmark against\n  Opus 5's 73.1%. Right for decisions that are gated and reversible; wrong for\n  final answers.\n\nQuick Start\nGet an API key\n\nCreate one at console.typesafe.ai/keys.\nConfigure\nUse it\n\ntryDecide() returns null on any failure. Use decide() when you want the\nfailure to surface; it throws a ProviderError whose cause carries a typed\nkind.\n\nFrom the CLI\n\nneurolink decide [state] calls decide() under the hood and reads\nTYPESAFE_API_KEY from the environment like every other CLI command. See the\nCLI command reference.\n\nThe degradation contract\n\nSetting the key is the entire switch for TypeSafe, and removing it is a\ncomplete undo while no other decision provider is configured. Every internal\nconsumer of decide fails open: with no decision provider configured, model\nrouting, context budgeting, relevance compaction, tool routing and RAG planning\nall behave exactly as they did before. Note that Perplexity's key is shared with\nits text provider, so a PERPLEXITY_API_KEY set for that provider also counts as\na configured decision provider; see\nOne key, two providers. The same\ngoes for CLOUDFLARE_API_KEY with CLOUDFLARE_ACCOUNT_ID, which the Workers AI\ntext provider reads too: set together they configure Cloudflare Clef, which ranks\nafter TypeSafe, Laya, XOR, SubQ and Perplexity; see One token, two providers.\nOPENAI_API_KEY, which the OpenAI chat provider reads too, configures OpenAI\nDecisions, which ranks after all of them but ScaleDown; see\nOne key, two providers.\nSCALEDOWN_API_KEY is ScaleDown's own and configures only it, as the last of\nthe eight; see ScaleDown. There is no\nconfiguration in which a missing, invalid, slow or unreachable decision model\nchanges NeuroLink's observable behaviour.\n\nA credential the service does not accept disables that provider instance rather\nthan paying a round trip on every later call to be told so again.\n\nTwo transports\n\nThe same model is reachable two ways, and the choice is made once in the\nconstructor.\n\n|                     | Direct              | Vercel AI Gateway                             |\n| ------------------- | ------------------- | --------------------------------------------- |\n| Key                 | TYPESAFE_API_KEY  | AI_GATEWAY_API_KEY                          |\n| Endpoint            | api.typesafe.ai   | ai-gateway.vercel.sh/v4/ai/evaluation-model |\n| Endpoint override   | TYPESAFE_BASE_URL | TYPESAFE_GATEWAY_URL                        |\n| Model named in      | request body        | ai-model-id header                          |\n| Question vocabulary | noul              | boolean                                     |\n| confidence        | on each answer      | on providerMetadata                         |\n| Billed by           | TypeSafe            | Vercel                                        |\n\nHolding both keys keeps the direct tra","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"","lvl3":""}},
+{"objectID":"7320dc5177f0e320fe7887f77c01c06519bb6c0ddeb8a8b38ecfa013a9363a1f","title":"TypeSafe (Jev) Provider Guide","url":"/docs/getting-started/providers/typesafe#typesafe-jev-provider-guide","content":"One of the providers that serve decide rather than generate/stream —\nit returns typed, calibrated judgments and emits no text at all. The others are\nLaya and XOR, open-weights models served from an endpoint\nyou configure, and SubQ, Perplexity,\nCloudflare Clef, OpenAI Decisions\nand ScaleDown, hosted APIs. When more than one is configured, TypeSafe is tried before the\nothers.","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"TypeSafe (Jev) Provider Guide","lvl3":""}},
 {"objectID":"8cc640004e04379192695b8fcd3d49552ac0f4ebd763093bc610b4c93e823fee","title":"Overview","url":"/docs/getting-started/providers/typesafe#overview","content":"TypeSafe's Jev is a \"System One\" model. You send one state plus a map of\nnamed, typed questions; it returns one typed answer per question, all evaluated\nin a single parallel pass. Nothing has to be parsed back out of prose, and every\nchoice/score answer carries a calibrated confidence rather than a\nself-reported one.\n\nBecause it emits no text, generate() and stream() are not available and\ngetAISDKModel() throws — the same shape Voyage and Jina already use for\nembedding-only providers. Its descriptor declares inferenceKinds: [\"decide\"],\nwhich keeps it out of auto-select and the health sweep, so those throws are\nunreachable in normal use.\n\nThis is not neurolink.evaluate(), which\nscores an already-generated response with RAGAS scorers. Different feature,\ndifferent word.","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"Overview","lvl3":""}},
-{"objectID":"4921fe7ae1c50c817be3b7742587c455084bee9d988015cf475edd5c3c196b84","title":"Key Facts","url":"/docs/getting-started/providers/typesafe#key-facts","content":"Provider id: typesafe (aliases: jev, typesafe-ai)\nInference kinds: decide only — one of the providers that do (the others\n  are Laya, XOR, SubQ,\n  Perplexity,\n  Cloudflare Clef and\n  OpenAI Decisions); TypeSafe is tried before them when\n  more than one is configured\nTool calling: none (toolSupport: \"none\") — a decision model calls nothing\nHealth check: env-only; it is never probed with a live generation\nDefault decide timeout: 5000 ms (timeouts.decideMs)\nLatency: flat in question count — 1 question ~393 ms, 400 questions\n  ~465 ms. Concurrent requests queue instead, so batch every question into one\n  call rather than fanning out.\nCost: ~$0.042 per million input tokens, output billed at zero — about\n  $0.00002 per decision. Output tokens are reported — measured 21 for a\n  single question, converging to ~17.5 per question in a batch of eight — they\n  are simply not charged.\nAccuracy is the trade: 67.8% on TypeSafe's own 711-case benchmark against\n  Opus 5's 73.1%. Right for decisions that are gated and reversible; wrong for\n  final answers.","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"Key Facts","lvl3":""}},
+{"objectID":"4921fe7ae1c50c817be3b7742587c455084bee9d988015cf475edd5c3c196b84","title":"Key Facts","url":"/docs/getting-started/providers/typesafe#key-facts","content":"Provider id: typesafe (aliases: jev, typesafe-ai)\nInference kinds: decide only — one of the providers that do (the others\n  are Laya, XOR, SubQ,\n  Perplexity,\n  Cloudflare Clef,\n  OpenAI Decisions and\n  ScaleDown); TypeSafe is tried before them when\n  more than one is configured\nTool calling: none (toolSupport: \"none\") — a decision model calls nothing\nHealth check: env-only; it is never probed with a live generation\nDefault decide timeout: 5000 ms (timeouts.decideMs)\nLatency: flat in question count — 1 question ~393 ms, 400 questions\n  ~465 ms. Concurrent requests queue instead, so batch every question into one\n  call rather than fanning out.\nCost: ~$0.042 per million input tokens, output billed at zero — about\n  $0.00002 per decision. Output tokens are reported — measured 21 for a\n  single question, converging to ~17.5 per question in a batch of eight — they\n  are simply not charged.\nAccuracy is the trade: 67.8% on TypeSafe's own 711-case benchmark against\n  Opus 5's 73.1%. Right for decisions that are gated and reversible; wrong for\n  final answers.","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"Key Facts","lvl3":""}},
 {"objectID":"f5f7265f7ee9ba7ead6f6016a4b9232258ffaf4256e544874d6bcca8323dbb04","title":"1. Get an API key","url":"/docs/getting-started/providers/typesafe#1-get-an-api-key","content":"Create one at console.typesafe.ai/keys.","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"1. Get an API key","lvl3":""}},
 {"objectID":"a7cfa6599d9458e8c4a9bfcef1bba07a4948ebe6b0f30a1352df07a713040d8e","title":"3. Use it","url":"/docs/getting-started/providers/typesafe#3-use-it","content":"tryDecide() returns null on any failure. Use decide() when you want the\nfailure to surface; it throws a ProviderError whose cause carries a typed\nkind.","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"3. Use it","lvl3":""}},
 {"objectID":"4ec7febf2956a104b0037472bab2e6c652e17df7f7cfdf69b73ebe2c5003f082","title":"From the CLI","url":"/docs/getting-started/providers/typesafe#from-the-cli","content":"neurolink decide [state] calls decide() under the hood and reads\nTYPESAFE_API_KEY from the environment like every other CLI command. See the\nCLI command reference.","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"From the CLI","lvl3":""}},
-{"objectID":"98aafdb50791dab3469be7930897496cf4d2fb9dc57e42bacb2ee179ac0064f2","title":"The degradation contract","url":"/docs/getting-started/providers/typesafe#the-degradation-contract","content":"Setting the key is the entire switch for TypeSafe, and removing it is a\ncomplete undo while no other decision provider is configured. Every internal\nconsumer of decide fails open: with no decision provider configured, model\nrouting, context budgeting, relevance compaction, tool routing and RAG planning\nall behave exactly as they did before. Note that Perplexity's key is shared with\nits text provider, so a PERPLEXITY_API_KEY set for that provider also counts as\na configured decision provider; see\nOne key, two providers. The same\ngoes for CLOUDFLARE_API_KEY with CLOUDFLARE_ACCOUNT_ID, which the Workers AI\ntext provider reads too: set together they configure Cloudflare Clef, which ranks\nafter TypeSafe, Laya, XOR, SubQ and Perplexity; see One token, two providers.\nOPENAI_API_KEY, which the OpenAI chat provider reads too, configures OpenAI\nDecisions, the last fallback; see\nOne key, two providers. There is no\nconfiguration in which a missing, invalid, slow or unreachable decision model\nchanges NeuroLink's observable behaviour.\n\nA credential the service does not accept disables that provider instance rather\nthan paying a round trip on every later call to be told so again.","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"The degradation contract","lvl3":""}},
+{"objectID":"98aafdb50791dab3469be7930897496cf4d2fb9dc57e42bacb2ee179ac0064f2","title":"The degradation contract","url":"/docs/getting-started/providers/typesafe#the-degradation-contract","content":"Setting the key is the entire switch for TypeSafe, and removing it is a\ncomplete undo while no other decision provider is configured. Every internal\nconsumer of decide fails open: with no decision provider configured, model\nrouting, context budgeting, relevance compaction, tool routing and RAG planning\nall behave exactly as they did before. Note that Perplexity's key is shared with\nits text provider, so a PERPLEXITY_API_KEY set for that provider also counts as\na configured decision provider; see\nOne key, two providers. The same\ngoes for CLOUDFLARE_API_KEY with CLOUDFLARE_ACCOUNT_ID, which the Workers AI\ntext provider reads too: set together they configure Cloudflare Clef, which ranks\nafter TypeSafe, Laya, XOR, SubQ and Perplexity; see One token, two providers.\nOPENAI_API_KEY, which the OpenAI chat provider reads too, configures OpenAI\nDecisions, which ranks after all of them but ScaleDown; see\nOne key, two providers.\nSCALEDOWN_API_KEY is ScaleDown's own and configures only it, as the last of\nthe eight; see ScaleDown. There is no\nconfiguration in which a missing, invalid, slow or unreachable decision model\nchanges NeuroLink's observable behaviour.\n\nA credential the service does not accept disables that provider instance rather\nthan paying a round trip on every later call to be told so again.","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"The degradation contract","lvl3":""}},
 {"objectID":"8fd6c7607ca461fdf5e7fb3b743b1761645c1ca9a2e666330b59614acb6455fa","title":"Two transports","url":"/docs/getting-started/providers/typesafe#two-transports","content":"The same model is reachable two ways, and the choice is made once in the\nconstructor.\n\n|                     | Direct              | Vercel AI Gateway                             |\n| ------------------- | ------------------- | --------------------------------------------- |\n| Key                 | TYPESAFE_API_KEY  | AI_GATEWAY_API_KEY                          |\n| Endpoint            | api.typesafe.ai   | ai-gateway.vercel.sh/v4/ai/evaluation-model |\n| Endpoint override   | TYPESAFE_BASE_URL | TYPESAFE_GATEWAY_URL                        |\n| Model named in      | request body        | ai-model-id header                          |\n| Question vocabulary | noul              | boolean                                     |\n| confidence        | on each answer      | on providerMetadata                         |\n| Billed by           | TypeSafe            | Vercel                                        |\n\nHolding both keys keeps the direct transport, so the confidence figures a\nhost already sees do not shift underneath it when a second key appears. Force\none with TYPESAFE_TRANSPORT=direct|gateway or\ncredentials.typesafe.transport. Either endpoint can be moved without a\nrelease: credentials.typesafe.baseURL / TYPESAFE_BASE_URL for the direct\none, credentials.typesafe.gatewayURL / TYPESAFE_GATEWAY_URL for the\ngateway route.\n\n⚠️ The gateway refuses every request — free credits included — until the\nVercel team has a credit card on file, returning 403\ncustomer_verification_required. That is an account state, not a bad key, and it\narrives before the model id is validated.\n\nFull detail, including the measured error table and why the distribution peak is\nnot a substitute for the reported confidence, is in\nThe decide inference type.","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"Two transports","lvl3":""}},
 {"objectID":"ac6585086d9cf682e7b1a66265cacfd1c4e961588384fe3883636c5deb66588c","title":"What NeuroLink uses it for","url":"/docs/getting-started/providers/typesafe#what-neurolink-uses-it-for","content":"| Area                                                             | What the decision replaces                                      |\n| ---------------------------------------------------------------- | --------------------------------------------------------------- |\n| Model routing   | difficulty + capabilities + risk + model pick in one round trip |\n| Model catalogue      | one choice over the registry ranks all N candidates at once   |\n| Context budget                  | a rubric-placed scope reading lowers the compaction threshold   |\n| Relevance compaction      | per-message keep/drop, plus a gate on the generated summary     |\n| Tool / MCP routing | one boolean per server, replacing a 15s LLM call at ~400 ms   |\n| RAG retrieval           | per-query topK / hybrid / graph / rerank planning             |","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"What NeuroLink uses it for","lvl3":""}},
 {"objectID":"04bf675bad2aeefe749c5229a396a476273c9b463d3b81a6767b9ca9357aa36d","title":"Limits and gotchas","url":"/docs/getting-started/providers/typesafe#limits-and-gotchas","content":"Two input ceilings, both enforced by the service: state plus the longest\n  single question ≈ 33,000 tokens, and state plus all questions ≈\n  64,000. Exceeding either returns max_tokens_exceeded with no message\n  at all — the provider supplies a real sentence in its place.\nBatch, never fan out. Latency is flat in question count but concurrent\n  requests queue, so a second round trip costs far more than a hundred extra\n  questions.\nA boolean carries no confidence of its own. Use\n  decisionBooleanConfidence(p) — distance from a coin flip, so 0.5 → 0 and\n  0/1 → 1.\n403 vs 401 are inverted on the direct API, and from TypeSafe's own docs: a\n  missing Authorization header returns 403, an invalid key returns\nThe gateway does not share this quirk.","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"Limits and gotchas","lvl3":""}},
 {"objectID":"7d7786b2749554b22a763088bb2c011be0b62bdeb53ee881e9331cf8616dffa7","title":"Troubleshooting","url":"/docs/getting-started/providers/typesafe#troubleshooting","content":"| Symptom                              | Cause                                                                 | Fix                                                                           |\n| ------------------------------------ | --------------------------------------------------------------------- | ----------------------------------------------------------------------------- |\n| tryDecide() always returns null  | No key, or the key was rejected once and the instance disabled itself | Check TYPESAFE_API_KEY; construct a new instance after fixing it            |\n| max_tokens_exceeded                | One of the two input ceilings                                         | Shorten state, or split questions across calls — but prefer shrinking state |\n| 403 customer_verification_required | Gateway transport, no card on the Vercel team                         | Add a payment method, or use the direct transport                             |\n| Routing never changes                | A modelPool is configured, which owns selection outright            | See Provider Orchestration           |","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"Troubleshooting","lvl3":""}},
-{"objectID":"6116c23ca86f351b03ac4844a1d76d9de849fb522ab4a03eb37f9e75017e567c","title":"See also","url":"/docs/getting-started/providers/typesafe#see-also","content":"The decide inference type — the full reference\nModel routing with a decision model\nLaya Provider Guide\nXOR Provider Guide\nPerplexity Decisions Provider Guide\nCloudflare Clef Provider Guide\nSubQ Provider Guide\nOpenAI Decisions Provider Guide\nProvider setup overview","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"See also","lvl3":""}},
+{"objectID":"6116c23ca86f351b03ac4844a1d76d9de849fb522ab4a03eb37f9e75017e567c","title":"See also","url":"/docs/getting-started/providers/typesafe#see-also","content":"The decide inference type — the full reference\nModel routing with a decision model\nLaya Provider Guide\nXOR Provider Guide\nPerplexity Decisions Provider Guide\nCloudflare Clef Provider Guide\nSubQ Provider Guide\nOpenAI Decisions Provider Guide\nScaleDown Provider Guide\nProvider setup overview","hierarchy":{"lvl0":"Getting Started","lvl1":"TypeSafe (Jev) Provider Guide","lvl2":"See also","lvl3":""}},
 {"objectID":"0db36dc931d978469a7ad4e6468f4b52f77cc7c0db36272c6cfe8ac11e411f42","title":"Umans AI Provider Guide","url":"/docs/getting-started/providers/umans-ai","content":"Umans AI Provider Guide\n\nUmans AI is a Tier-2 catalog provider: OpenAI-wire-compatible with no\nbehavioural quirks, so its entire integration is one JSON file\n(src/lib/providers/catalog/umans-ai.json) rather than hand-written code.\nThat file is the source of truth for everything on this page.\n\nVerification status: this entry is docs- and roster-verified, not yet\nlive-verified. Every field below comes from Umans AI's own public pages and\nunauthenticated GET requests — no account was created and no API key was\nused to build it. evidence.liveMatrix is null until someone runs the live\ncapability matrix with a real key (see\nVerification status below).\n\nKey Facts\nProvider id: umans-ai (alias umans)\nProtocol: OpenAI-compatible (/chat/completions) — the docs say \"Umans\n  Code also implements the OpenAI Chat Completions API\"\n  (docs). The same gateway also\n  documents an Anthropic Messages route (POST /v1/messages) and a Responses\n  route (POST /v1/responses); this entry uses only /chat/completions\nBase URL: https://api.code.umans.ai/v1\nDefault model: umans-coder — a routing alias, not a fixed model, see\n  Models\nModels in catalog: 7 (curated from a 10-model live roster; three ids are\n  left out on purpose — umans-deepseek-v4.1-flash-lab and\n  umans-mimo-v2.6-pro-lab, which /v1/models/info describes as Labs\n  experiments, \"temporary, not a permanent id\", with access \"seat-gated through\n  the Labs page\", and umans-qwen3.6-35b-a3b, which it describes as a\n  \"Technical alias for umans-flash\")\nStreaming: supported — the docs' OpenAI example sends \"stream\": true to\n  POST /v1/chat/completions, and the reasoning section says reasoning \"streams\n  live, token by token, alongside the answer\"\nTool calling: true — GET /v1/models/info reports supports_tools: true\n  for every roster model. The docs show no tools request example for\n  /v1/chat/completions, so this rests on that field alone and has not been\n  exercised\nTools while streaming: not declared (false) — the docs show no\n  tool-call request or response on this route, streaming or otherwise\nStructured output: not declared (false) — no page opened mentions\n  response_format, json_schema or json_object\nStructured output + tools together: not declared (false) — no combined\n  probe was run without credentials\nEmbeddings: not declared — the docs describe no embeddings route\nThinking: true — the docs' \"Reasoning & Extended Thinking\" section\n  documents reasoning_effort on /v1/chat/completions and returns reasoning\n  in a reasoning_content field, separate from content. Which levels a model\n  accepts varies, see Models\nBilling: no-free-tier — a prepaid wallet billed per token, see\n  Billing below\nKey format: the Umans pages conflict, so apiKeyFormat is null and\n  NeuroLink does not validate a prefix. The docs' \"Any BYOK Tool\" section says\n  the key \"starts with sk-\" and their examples write it as\n  sk-your-umans-api-key, while the docs' wallet-summary example shows a key\n  prefix of umans_7fKq and the homepage .env example shows umans_\n\nQuick Start\nGet an API key\nVisit: https://app.umans.ai/register and create an account with Google or an email and password (already have one? sign in at https://app.umans.ai/login). Organizations are provisioned by Umans: email contact@umans.ai (https://app.umans.ai/offers/code/docs/orgs)\nOpen https://app.umans.ai/billing, go to Dashboard -> API Keys and generate a key; the docs say it is shown only once, so copy it immediately\nPrepaid wallet, no free tier stated: top up before calling the API - the docs say a wallet with no credit (no paid top-up and no active promo grant) \"does not serve until it has some\", and promo credit serves only while it lasts (https://app.umans.ai/offers/code/docs)\nSet UMANS_AI_API_KEY in your .env file\nConfigure\nUse it\n\nPer-request credentials work as they do for every provider:\n\nBilling\n\nUmans AI's pricing page (https://app.umans.ai/pricing) says \"Pay per token. Top\nup, create a key, and go.\" and \"every request debits exactly what the model\nused\". The docs (https://app.umans.ai/offers/code/docs) say a wallet with no\ncredit at all — no paid top-up and no active promo grant — \"does not serve until\nit has some\"; promo credit \"serves while it lasts\" and then the wallet suspends\nuntil a paid top-up. A funded wallet that drains to its tier's floor (a small\nnegative allowance, $5 at Tiers 0-1) also suspends, and the docs call that \"a\nstop, not a 429\". No page states a free tier or standing free credits, so the\nentry records no-free-tier. The login page's \"Sign up for free\" link is about\ncreating an account; it is not a stated free allowance. The docs document\nGET /v1/usage and a wallet-summary route on app.umans.ai for checking\nbalance and spend. Both need a key; /v1/usage was called only without one,\nfor the 401 probe under Verification status.\n\nRequest limits sit on top of the balance. The docs give each wallet tier a\nrolling five-hour request window and a concurrency cap: Tier 0 (first top-up)\n2,000 requests and 4 in flight","hierarchy":{"lvl0":"Getting Started","lvl1":"Umans AI Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"ff607cfd31ffc8da0d09dd8b747b2825462342a8683ab3c6cd35251b2538b0ef","title":"Umans AI Provider Guide","url":"/docs/getting-started/providers/umans-ai#umans-ai-provider-guide","content":"Umans AI is a Tier-2 catalog provider: OpenAI-wire-compatible with no\nbehavioural quirks, so its entire integration is one JSON file\n(src/lib/providers/catalog/umans-ai.json) rather than hand-written code.\nThat file is the source of truth for everything on this page.\n\nVerification status: this entry is docs- and roster-verified, not yet\nlive-verified. Every field below comes from Umans AI's own public pages and\nunauthenticated GET requests — no account was created and no API key was\nused to build it. evidence.liveMatrix is null until someone runs the live\ncapability matrix with a real key (see\nVerification status below).","hierarchy":{"lvl0":"Getting Started","lvl1":"Umans AI Provider Guide","lvl2":"Umans AI Provider Guide","lvl3":""}},
 {"objectID":"b71e70c479a72c1b2c02508e7f57c9dc8f5bd366f7dab23c56b323d7dbb8e3ca","title":"Key Facts","url":"/docs/getting-started/providers/umans-ai#key-facts","content":"Provider id: umans-ai (alias umans)\nProtocol: OpenAI-compatible (/chat/completions) — the docs say \"Umans\n  Code also implements the OpenAI Chat Completions API\"\n  (docs). The same gateway also\n  documents an Anthropic Messages route (POST /v1/messages) and a Responses\n  route (POST /v1/responses); this entry uses only /chat/completions\nBase URL: https://api.code.umans.ai/v1\nDefault model: umans-coder — a routing alias, not a fixed model, see\n  Models\nModels in catalog: 7 (curated from a 10-model live roster; three ids are\n  left out on purpose — umans-deepseek-v4.1-flash-lab and\n  umans-mimo-v2.6-pro-lab, which /v1/models/info describes as Labs\n  experiments, \"temporary, not a permanent id\", with access \"seat-gated through\n  the Labs page\", and umans-qwen3.6-35b-a3b, which it describes as a\n  \"Technical alias for umans-flash\")\nStreaming: supported — the docs' OpenAI example sends \"stream\": true to\n  POST /v1/chat/completions, and the reasoning section says reasoning \"streams\n  live, token by token, alongside the answer\"\nTool calling: true — GET /v1/models/info reports supports_tools: true\n  for every roster model. The docs show no tools request example for\n  /v1/chat/completions, so this rests on that field alone and has not been\n  exercised\nTools while streaming: not declared (false) — the docs show no\n  tool-call request or response on this route, streaming or otherwise\nStructured output: not declared (false) — no page opened mentions\n  response_format, json_schema or json_object\nStructured output + tools together: not declared (false) — no combined\n  probe was run without credentials\nEmbeddings: not declared — the docs describe no embeddings route\nThinking: true — the docs' \"Reasoning & Extended Thinking\" section\n  documents reasoning_effort on /v1/chat/completions and returns reasoning\n  in a reasoning_content field, separate from content.","hierarchy":{"lvl0":"Getting Started","lvl1":"Umans AI Provider Guide","lvl2":"Key Facts","lvl3":""}},
@@ -7485,19 +7513,19 @@
 {"objectID":"7bc31a9a4e8ceaa9d384ce6953e41c6991344d5affeecf269a14248d769eb2f3","title":"\"xAI account has insufficient quota\"","url":"/docs/getting-started/providers/xai#xai-account-has-insufficient-quota","content":"Top up at https://console.x.ai/.","hierarchy":{"lvl0":"Getting Started","lvl1":"xAI Grok Provider Guide","lvl2":"\"xAI account has insufficient quota\"","lvl3":""}},
 {"objectID":"62c0c746ffa186a922276c344852f82e4ee3242f81c67b58a433defc4f4a4924","title":"\"Model not found\"","url":"/docs/getting-started/providers/xai#model-not-found","content":"Use one of the documented model IDs above. Custom fine-tunes are not\nexposed through the public API at this time.","hierarchy":{"lvl0":"Getting Started","lvl1":"xAI Grok Provider Guide","lvl2":"\"Model not found\"","lvl3":""}},
 {"objectID":"a55b6d892c11d17b5222d76096529452815cd15809b93c491a3c484756185910","title":"See Also","url":"/docs/getting-started/providers/xai#see-also","content":"Adding a new LLM provider — internal reference for the integration pattern this provider follows\nDeepSeek Provider — sibling OpenAI-compat provider with reasoning models\nGroq Provider — sibling OpenAI-compat provider with sub-100ms inference\n\nNeed Help? Open a GitHub Discussion or issue.","hierarchy":{"lvl0":"Getting Started","lvl1":"xAI Grok Provider Guide","lvl2":"See Also","lvl3":""}},
-{"objectID":"03a891da2cc3529b3e4bfc042e5834454a779a2d3fa9389367ba4b3d05266871","title":"XOR Provider Guide","url":"/docs/getting-started/providers/xor","content":"XOR Provider Guide\n\nA provider of decide — the same typed boolean / choice /\nscore answers as TypeSafe's Jev and Laya, from\nJuspay's open-weights model, which also reads images and a video. It emits no\ntext at all.\n\nOverview\n\nXOR is Juspay's Apache-2.0 decision model, post-trained from Qwen3.6-35B-A3B;\nits model id is xor-1.1. You send one state plus named, typed questions, and\noptionally images or a video. XOR answers every question in a single batched\npass and returns a typed answer for each. NeuroLink calls whatever base URL you\nconfigure — a deployment of the model, or a LiteLLM proxy with a route to one.\nThere is no built-in endpoint. The weights and setup instructions are at\nhuggingface.co/juspay/xor.\n\nKey Facts\n\n|                       |                                                                      |\n| --------------------- | -------------------------------------------------------------------- |\n| Inference type        | decide only                                                        |\n| Model                 | xor-1.1 by default; XOR_MODEL sets the name your proxy uses      |\n| Media                 | up to 8 images and one video; the whole request body at most 8 MB    |\n| Input window          | about 200,000 estimated tokens of state                              |\n| Questions per request | no cap in NeuroLink; a choice or score takes 2 to 255 options    |\n| Endpoint              | <base URL>/v1/systemone; the base URL is required, with no default |\n| Precedence            | used automatically only when TypeSafe and Laya are not configured    |\n\nQuick Start\nGet an endpoint and a key\n\nRun a deployment of XOR (see\nhuggingface.co/juspay/xor), or use a\nLiteLLM proxy with a route to one. NeuroLink posts to <base URL>/v1/systemone.\nThe base URL is the origin of the deployment or of the proxy route, and a\ntrailing /v1 is accepted. It must be https://; http:// is accepted only for\nlocalhost, 127.0.0.1 and [::1], because every request carries the key and the\nstate. A deployment reachable only over plain http inside a network needs an\nhttps:// front, or a port-forward to localhost. On a LiteLLM proxy the key's team must be allowed\nthe model you ask for (xor-1.1 by default); otherwise the proxy answers 403\nteam_model_access_denied. A proxy may serve XOR under a name of its own: set\nXOR_MODEL to that name.\nConfigure\n\nBoth the base URL and the key are required. Set them in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:\nUse it\n\nFrom the CLI\n\nThe CLI reads the same environment variables:\n\nImages and video\n\nXOR reads images and one video alongside the state. TypeSafe and Laya do not\nread media, and Perplexity,\nCloudflare Clef and OpenAI Decisions\nread images but no video.\n\nFrom the CLI, pass --image <path> (repeatable) and --video <path>:\n\nThe rules:\nUp to 8 images and one video per request, in images and video.\nEach is a Buffer, a local file path or a data: URL. An http(s) URL is\n  refused: NeuroLink does not fetch media for you.\nThe type comes from the bytes, not the file extension: PNG, JPEG, WebP\n  and GIF images, and MP4, MOV and WebM video. NeuroLink\n  sends a Buffer or a file to XOR as a data: URL. A data: URL you pass\n  yourself must hold base64 image or video content, and is sent as given.\nThe whole request body may be at most 8 MB. The limit applies to the\n  encoded body, so the base64 form counts, not the size of the files on disk. A\n  file over the limit is refused from its size, before it is read.\nWhat can be checked locally is refused before any request, as a\n  non-retryable invalid_request: a missing file, a directory, an empty Buffer\n  or file, bytes that are not a recognised image or video, a remote URL or any\n  other URL scheme, a string that is not a path or a data: URL, more than 8\n  images, or a body over 8 MB. Bytes with the right signature that the model\n  cannot decode are sent, and the server's refusal comes back as a server\n  error, retried once.\nTypeSafe and Laya refuse media too, before any request, and the error\n  names the providers that accept it. Perplexity, Cloudflare Clef and OpenAI\n  Decisions refuse a video.\nImages and a video can be sent together (up to 8 images and one video),\n  but the model does not reliably tell the two apart.\nThe result carries mediaBytes, the encoded size of the media sent. The\n  decide spans carry decision.images.count and decision.media.bytes, and\n  never any base64.\n\nWhen NeuroLink uses it\n\nEvery built-in consumer of decide — model routing, relevance-driven\ncompaction, tool routing and RAG planning — asks for the default decision\nprovider. That is the first one that is configured, in the environment or in the\ncredentials passed to the SDK, in the order TypeSafe, Laya, XOR,\nSubQ, Perplexity, Cloudflare Clef, then\nOpenAI Decisions. TypeSafe counts with either of its keys,\nTYPESAFE_API_KEY or AI_GATEWAY_API_KEY. Laya and X","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"","lvl3":""}},
+{"objectID":"03a891da2cc3529b3e4bfc042e5834454a779a2d3fa9389367ba4b3d05266871","title":"XOR Provider Guide","url":"/docs/getting-started/providers/xor","content":"XOR Provider Guide\n\nA provider of decide — the same typed boolean / choice /\nscore answers as TypeSafe's Jev and Laya, from\nJuspay's open-weights model, which also reads images and a video. It emits no\ntext at all.\n\nOverview\n\nXOR is Juspay's Apache-2.0 decision model, post-trained from Qwen3.6-35B-A3B;\nits model id is xor-1.1. You send one state plus named, typed questions, and\noptionally images or a video. XOR answers every question in a single batched\npass and returns a typed answer for each. NeuroLink calls whatever base URL you\nconfigure — a deployment of the model, or a LiteLLM proxy with a route to one.\nThere is no built-in endpoint. The weights and setup instructions are at\nhuggingface.co/juspay/xor.\n\nKey Facts\n\n|                       |                                                                      |\n| --------------------- | -------------------------------------------------------------------- |\n| Inference type        | decide only                                                        |\n| Model                 | xor-1.1 by default; XOR_MODEL sets the name your proxy uses      |\n| Media                 | up to 8 images and one video; the whole request body at most 8 MB    |\n| Input window          | about 200,000 estimated tokens of state                              |\n| Questions per request | no cap in NeuroLink; a choice or score takes 2 to 255 options    |\n| Endpoint              | <base URL>/v1/systemone; the base URL is required, with no default |\n| Precedence            | used automatically only when TypeSafe and Laya are not configured    |\n\nQuick Start\nGet an endpoint and a key\n\nRun a deployment of XOR (see\nhuggingface.co/juspay/xor), or use a\nLiteLLM proxy with a route to one. NeuroLink posts to <base URL>/v1/systemone.\nThe base URL is the origin of the deployment or of the proxy route, and a\ntrailing /v1 is accepted. It must be https://; http:// is accepted only for\nlocalhost, 127.0.0.1 and [::1], because every request carries the key and the\nstate. A deployment reachable only over plain http inside a network needs an\nhttps:// front, or a port-forward to localhost. On a LiteLLM proxy the key's team must be allowed\nthe model you ask for (xor-1.1 by default); otherwise the proxy answers 403\nteam_model_access_denied. A proxy may serve XOR under a name of its own: set\nXOR_MODEL to that name.\nConfigure\n\nBoth the base URL and the key are required. Set them in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:\nUse it\n\nFrom the CLI\n\nThe CLI reads the same environment variables:\n\nImages and video\n\nXOR reads images and one video alongside the state. TypeSafe, Laya,\nSubQ and ScaleDown do not read media, and Perplexity,\nCloudflare Clef and OpenAI Decisions\nread images but no video.\n\nFrom the CLI, pass --image <path> (repeatable) and --video <path>:\n\nThe rules:\nUp to 8 images and one video per request, in images and video.\nEach is a Buffer, a local file path or a data: URL. An http(s) URL is\n  refused: NeuroLink does not fetch media for you.\nThe type comes from the bytes, not the file extension: PNG, JPEG, WebP\n  and GIF images, and MP4, MOV and WebM video. NeuroLink\n  sends a Buffer or a file to XOR as a data: URL. A data: URL you pass\n  yourself must hold base64 image or video content, and is sent as given.\nThe whole request body may be at most 8 MB. The limit applies to the\n  encoded body, so the base64 form counts, not the size of the files on disk. A\n  file over the limit is refused from its size, before it is read.\nWhat can be checked locally is refused before any request, as a\n  non-retryable invalid_request: a missing file, a directory, an empty Buffer\n  or file, bytes that are not a recognised image or video, a remote URL or any\n  other URL scheme, a string that is not a path or a data: URL, more than 8\n  images, or a body over 8 MB. Bytes with the right signature that the model\n  cannot decode are sent, and the server's refusal comes back as a server\n  error, retried once.\nTypeSafe and Laya refuse media too, before any request, and the error\n  names the providers that accept it. Perplexity, Cloudflare Clef and OpenAI\n  Decisions refuse a video.\nImages and a video can be sent together (up to 8 images and one video),\n  but the model does not reliably tell the two apart.\nThe result carries mediaBytes, the encoded size of the media sent. The\n  decide spans carry decision.images.count and decision.media.bytes, and\n  never any base64.\n\nWhen NeuroLink uses it\n\nEvery built-in consumer of decide — model routing, relevance-driven\ncompaction, tool routing and RAG planning — asks for the default decision\nprovider. That is the first one that is configured, in the environment or in the\ncredentials passed to the SDK, in the order TypeSafe, Laya, XOR,\nSubQ, Perplexity, Cloudflare Clef,\nOpenAI Decisions, then ScaleDown. TypeSafe counts with either of its keys,\nTYPESAFE_API_KEY or AI","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"5ed0d5759a64d045ba6c8acb04b0b6e6fef6f00399b88df9847c053d52a8e70c","title":"XOR Provider Guide","url":"/docs/getting-started/providers/xor#xor-provider-guide","content":"A provider of decide — the same typed boolean / choice /\nscore answers as TypeSafe's Jev and Laya, from\nJuspay's open-weights model, which also reads images and a video. It emits no\ntext at all.","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"XOR Provider Guide","lvl3":""}},
 {"objectID":"62f1da4b03072e86ddddfcfed9098a9efa475d349ea26f88c901347eca7fc77a","title":"Overview","url":"/docs/getting-started/providers/xor#overview","content":"XOR is Juspay's Apache-2.0 decision model, post-trained from Qwen3.6-35B-A3B;\nits model id is xor-1.1. You send one state plus named, typed questions, and\noptionally images or a video. XOR answers every question in a single batched\npass and returns a typed answer for each. NeuroLink calls whatever base URL you\nconfigure — a deployment of the model, or a LiteLLM proxy with a route to one.\nThere is no built-in endpoint. The weights and setup instructions are at\nhuggingface.co/juspay/xor.","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"Overview","lvl3":""}},
 {"objectID":"2a68b1796658b1e42d1fb20868e2f28c9b9b13e9e83ebb2ac4411de95d15b8e3","title":"Key Facts","url":"/docs/getting-started/providers/xor#key-facts","content":"|                       |                                                                      |\n| --------------------- | -------------------------------------------------------------------- |\n| Inference type        | decide only                                                        |\n| Model                 | xor-1.1 by default; XOR_MODEL sets the name your proxy uses      |\n| Media                 | up to 8 images and one video; the whole request body at most 8 MB    |\n| Input window          | about 200,000 estimated tokens of state                              |\n| Questions per request | no cap in NeuroLink; a choice or score takes 2 to 255 options    |\n| Endpoint              | <base URL>/v1/systemone; the base URL is required, with no default |\n| Precedence            | used automatically only when TypeSafe and Laya are not configured    |","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"Key Facts","lvl3":""}},
 {"objectID":"06cdd61cf369f5af93c8e03841be4ebddcc8e1fde819b193e35a9fabfa7db51a","title":"1. Get an endpoint and a key","url":"/docs/getting-started/providers/xor#1-get-an-endpoint-and-a-key","content":"Run a deployment of XOR (see\nhuggingface.co/juspay/xor), or use a\nLiteLLM proxy with a route to one. NeuroLink posts to <base URL>/v1/systemone.\nThe base URL is the origin of the deployment or of the proxy route, and a\ntrailing /v1 is accepted. It must be https://; http:// is accepted only for\nlocalhost, 127.0.0.1 and [::1], because every request carries the key and the\nstate. A deployment reachable only over plain http inside a network needs an\nhttps:// front, or a port-forward to localhost. On a LiteLLM proxy the key's team must be allowed\nthe model you ask for (xor-1.1 by default); otherwise the proxy answers 403\nteam_model_access_denied. A proxy may serve XOR under a name of its own: set\nXOR_MODEL to that name.","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"1. Get an endpoint and a key","lvl3":""}},
 {"objectID":"540db8a33720c64f9e3e59c312e3d3e2476799c726c6e5729729ecaf2bbefa05","title":"2. Configure","url":"/docs/getting-started/providers/xor#2-configure","content":"Both the base URL and the key are required. Set them in the environment:\n\nor in the config passed to the SDK, exactly as for any other provider. Values\npassed per call override the constructor's, which override the environment:","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"2. Configure","lvl3":""}},
 {"objectID":"3fd4afc9a2670046c937e06586a65b3b5eadb6de1560ee1e1a3abf5f05524876","title":"From the CLI","url":"/docs/getting-started/providers/xor#from-the-cli","content":"The CLI reads the same environment variables:","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"From the CLI","lvl3":""}},
-{"objectID":"90a0f1f65868c0d5b68bb8fb5828a32f704d4ed3a164081275a6837b65a8e2f7","title":"Images and video","url":"/docs/getting-started/providers/xor#images-and-video","content":"XOR reads images and one video alongside the state. TypeSafe and Laya do not\nread media, and Perplexity,\nCloudflare Clef and OpenAI Decisions\nread images but no video.\n\nFrom the CLI, pass --image <path> (repeatable) and --video <path>:\n\nThe rules:\nUp to 8 images and one video per request, in images and video.\nEach is a Buffer, a local file path or a data: URL. An http(s) URL is\n  refused: NeuroLink does not fetch media for you.\nThe type comes from the bytes, not the file extension: PNG, JPEG, WebP\n  and GIF images, and MP4, MOV and WebM video. NeuroLink\n  sends a Buffer or a file to XOR as a data: URL. A data: URL you pass\n  yourself must hold base64 image or video content, and is sent as given.\nThe whole request body may be at most 8 MB. The limit applies to the\n  encoded body, so the base64 form counts, not the size of the files on disk. A\n  file over the limit is refused from its size, before it is read.\nWhat can be checked locally is refused before any request, as a\n  non-retryable invalid_request: a missing file, a directory, an empty Buffer\n  or file, bytes that are not a recognised image or video, a remote URL or any\n  other URL scheme, a string that is not a path or a data: URL, more than 8\n  images, or a body over 8 MB. Bytes with the right signature that the model\n  cannot decode are sent, and the server's refusal comes back as a server\n  error, retried once.\nTypeSafe and Laya refuse media too, before any request, and the error\n  names the providers that accept it. Perplexity, Cloudflare Clef and OpenAI\n  Decisions refuse a video.\nImages and a video can be sent together (up to 8 images and one video),\n  but the model does not reliably tell the two apart.\nThe result carries mediaBytes, the encoded size of the media sent. The\n  decide spans carry decision.images.count and decision.media.bytes, and\n  never any base64.","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"Images and video","lvl3":""}},
-{"objectID":"4fea1eacec247a5a95e9c0d017549cb5233da507c593a841ce0e392eb2d11e2a","title":"When NeuroLink uses it","url":"/docs/getting-started/providers/xor#when-neurolink-uses-it","content":"Every built-in consumer of decide — model routing, relevance-driven\ncompaction, tool routing and RAG planning — asks for the default decision\nprovider. That is the first one that is configured, in the environment or in the\ncredentials passed to the SDK, in the order TypeSafe, Laya, XOR,\nSubQ, Perplexity, Cloudflare Clef, then\nOpenAI Decisions. TypeSafe counts with either of its keys,\nTYPESAFE_API_KEY or AI_GATEWAY_API_KEY. Laya and XOR each count only with\nboth their key and their base URL. SubQ counts with its key alone. Perplexity\ncounts with its key alone, which is shared with Perplexity's text provider. Cloudflare Clef counts only with both\nCLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID, the same two variables the\nWorkers AI text provider reads. OpenAI Decisions counts with its key alone,\nOPENAI_API_KEY, which the OpenAI chat provider reads too. A caller can always\nname XOR with provider: \"xor\". So:\nA TypeSafe key, or Laya's key and base URL, plus XOR's key and base URL:\n  built-in features use TypeSafe or Laya, in that order; XOR runs only where a\n  caller asks for provider: \"xor\".\nXOR's key and base URL, plus SUBQ_API_KEY: built-in features use XOR;\n  SubQ runs only where a caller asks for provider: \"subq\".\nXOR's key and base URL, plus a Perplexity key: built-in features use XOR;\n  Perplexity runs only where a caller asks for\n  provider: \"perplexity-decider\".\nXOR's key and base URL, plus CLOUDFLARE_API_KEY and\n  CLOUDFLARE_ACCOUNT_ID: built-in features use XOR; Cloudflare Clef runs\n  only where a caller asks for provider: \"cloudflare-clef\".\nOnly XOR's key and base URL: built-in features use XOR.\nAn XOR key with no base URL: XOR is not configured. Built-in features\n  ignore it, and provider: \"xor\" fails with XOR requires a base URL.\nNone of TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef or OpenAI\n  Decisions: everything behaves exactly as it did without a decision model.","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"When NeuroLink uses it","lvl3":""}},
+{"objectID":"90a0f1f65868c0d5b68bb8fb5828a32f704d4ed3a164081275a6837b65a8e2f7","title":"Images and video","url":"/docs/getting-started/providers/xor#images-and-video","content":"XOR reads images and one video alongside the state. TypeSafe, Laya,\nSubQ and ScaleDown do not read media, and Perplexity,\nCloudflare Clef and OpenAI Decisions\nread images but no video.\n\nFrom the CLI, pass --image <path> (repeatable) and --video <path>:\n\nThe rules:\nUp to 8 images and one video per request, in images and video.\nEach is a Buffer, a local file path or a data: URL. An http(s) URL is\n  refused: NeuroLink does not fetch media for you.\nThe type comes from the bytes, not the file extension: PNG, JPEG, WebP\n  and GIF images, and MP4, MOV and WebM video. NeuroLink\n  sends a Buffer or a file to XOR as a data: URL. A data: URL you pass\n  yourself must hold base64 image or video content, and is sent as given.\nThe whole request body may be at most 8 MB. The limit applies to the\n  encoded body, so the base64 form counts, not the size of the files on disk. A\n  file over the limit is refused from its size, before it is read.\nWhat can be checked locally is refused before any request, as a\n  non-retryable invalid_request: a missing file, a directory, an empty Buffer\n  or file, bytes that are not a recognised image or video, a remote URL or any\n  other URL scheme, a string that is not a path or a data: URL, more than 8\n  images, or a body over 8 MB. Bytes with the right signature that the model\n  cannot decode are sent, and the server's refusal comes back as a server\n  error, retried once.\nTypeSafe and Laya refuse media too, before any request, and the error\n  names the providers that accept it. Perplexity, Cloudflare Clef and OpenAI\n  Decisions refuse a video.\nImages and a video can be sent together (up to 8 images and one video),\n  but the model does not reliably tell the two apart.\nThe result carries mediaBytes, the encoded size of the media sent. The\n  decide spans carry decision.images.count and decision.media.bytes, and\n  never any base64.","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"Images and video","lvl3":""}},
+{"objectID":"4fea1eacec247a5a95e9c0d017549cb5233da507c593a841ce0e392eb2d11e2a","title":"When NeuroLink uses it","url":"/docs/getting-started/providers/xor#when-neurolink-uses-it","content":"Every built-in consumer of decide — model routing, relevance-driven\ncompaction, tool routing and RAG planning — asks for the default decision\nprovider. That is the first one that is configured, in the environment or in the\ncredentials passed to the SDK, in the order TypeSafe, Laya, XOR,\nSubQ, Perplexity, Cloudflare Clef,\nOpenAI Decisions, then ScaleDown. TypeSafe counts with either of its keys,\nTYPESAFE_API_KEY or AI_GATEWAY_API_KEY. Laya and XOR each count only with\nboth their key and their base URL. SubQ counts with its key alone. Perplexity\ncounts with its key alone, which is shared with Perplexity's text provider. Cloudflare Clef counts only with both\nCLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID, the same two variables the\nWorkers AI text provider reads. OpenAI Decisions counts with its key alone,\nOPENAI_API_KEY, which the OpenAI chat provider reads too. ScaleDown counts with\nits key alone, SCALEDOWN_API_KEY, which no other provider reads. A caller can\nalways name XOR with provider: \"xor\". So:\nA TypeSafe key, or Laya's key and base URL, plus XOR's key and base URL:\n  built-in features use TypeSafe or Laya, in that order; XOR runs only where a\n  caller asks for provider: \"xor\".\nXOR's key and base URL, plus SUBQ_API_KEY: built-in features use XOR;\n  SubQ runs only where a caller asks for provider: \"subq\".\nXOR's key and base URL, plus SCALEDOWN_API_KEY: built-in features use XOR;\n  ScaleDown runs only where a caller asks for provider: \"scaledown\".\nXOR's key and base URL, plus a Perplexity key: built-in features use XOR;\n  Perplexity runs only where a caller asks for\n  provider: \"perplexity-decider\".\nXOR's key and base URL, plus CLOUDFLARE_API_KEY and\n  CLOUDFLARE_ACCOUNT_ID: built-in features use XOR; Cloudflare Clef runs\n  only where a caller asks for provider: \"cloudflare-clef\".\nOnly XOR's key and base URL: built-in features use XOR.\nAn XOR key with no base URL: XOR is not configured.","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"When NeuroLink uses it","lvl3":""}},
 {"objectID":"20a784de9dff5b3afa1f5ee7c0c60e268a7274b7e0c2a447aca431642f8fcabc","title":"Limits","url":"/docs/getting-started/providers/xor#limits","content":"The window is large, but shared. NeuroLink allows about 200,000 estimated\ntokens of state, and refuses more before any network call, with\nmax_tokens_exceeded. The size is an estimate, not XOR's tokenizer: about four\ncharacters per token for ASCII text, and one token per character for other\nscripts, which errs toward refusing. The 200,000 figure is a conservative\ndefault under the deployment's 250,000-token prefill, and that prefill is shared\nby the state, the questions and any media.\n\nThe figures in this paragraph are measurements of one live LiteLLM route, not\nlimits of XOR: the model behind the route's name has not been confirmed as XOR,\nand its 262,144-token context length differs from the 250,000-token prefill\nabove. They were taken by sending requests to the route directly rather than\nthrough NeuroLink, which refuses a state of 1.2 million ASCII characters before\nany request, at about 315,000 estimated tokens. A state of about 215,000 tokens\n(1.2 million characters of English) was accepted, and one of about 430,000\ntokens was refused with a 500 whose text says the input is longer than the\nmodel's context length. The estimate is not exact in either direction. English\naveraged about 5.6 characters per token on the route, so the estimate\nover-counts it. Dense JSON averaged about 1.8, so the estimate under-counts it\nby more than half, and a dense state estimated under the limit can still be\nlonger than the server's context. A structured (non-string) state of 100,000\nChinese characters passed NeuroLink's local check and was refused by the\nserver, which counted it as 485,774 tokens.\n\nA state near the limit, or a lot of media, can therefore still be refused by the\nserver as too long. A 413, or a 5xx other than 503 whose message says the\ncontext was too long, arrives as max_tokens_exceeded and is not retried; every\nother reply is classified as in the Errors table below. Built-in consumers treat\na refusal as \"carry on as before\".\n\nA proxy may cap parallel requests.","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"Limits","lvl3":""}},
 {"objectID":"d3945a40a2095e352b8f5f71228a59bbcb1cb60823696dbe20f3fafffe6b7c29","title":"Errors","url":"/docs/getting-started/providers/xor#errors","content":"| Reply                                                        | Kind                                                      | Retried   |\n| ------------------------------------------------------------ | --------------------------------------------------------- | --------- |\n| 401                                                          | authentication — the provider instance stops retrying   | no        |\n| 403 / 402                                                    | invalid_request — fixable; the instance is not disabled | no        |\n| 413, or a 5xx other than 503 saying the context was exceeded | max_tokens_exceeded                                     | no        |\n| any other 4xx (for example 422, 404)                         | invalid_request                                         | no        |\n| 429                                                          | rate_limit                                              | yes, once |\n| 503                                                          | overloaded                                              | yes, once |\n| other 5xx                                                    | server                                                  | yes, once |\n| state over the window (local)                                | max_tokens_exceeded, with no network call               | no        |\n| unusable media (local)                                       | invalid_request, with no network call                   | no        |\n| no base URL configured (local)                               | invalid_request, with no network call                   | no        |\n\nA 403 or 402 is deliberately not authentication. On a LiteLLM proxy a 403\nmeans the key's team does not allow the model you asked for\n(team_model_access_denied), and\na 402 means the team has no budget. An admin can fix either, so the provider\ninstance keeps working once it is fixed and is not disabled.","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"Errors","lvl3":""}},
 {"objectID":"37466cda852e51cad898009d283cae74e18a20774a3ca0fad9f146e3efb8a5a9","title":"Troubleshooting","url":"/docs/getting-started/providers/xor#troubleshooting","content":"XOR requires a base URL — set XOR_BASE_URL, or pass\n  credentials.xor.baseURL. There is no default endpoint.\nThe XOR base URL must not carry credentials… — the base URL has a user\n  name, a password, a query string or a fragment. Set it to the origin only, for\n  example https://your-proxy.example.com, and give the key through\n  XOR_API_KEY or credentials.xor.apiKey.\nThe XOR base URL must start with https:// or http:// — the value has no\n  scheme (for example xor.internal:8080) or uses another scheme. Add\n  https://, or http:// for a server on this machine.\nThe XOR base URL must use https:// unless it points at this machine — the\n  value is an http:// URL whose host is not localhost, 127.0.0.1 or [::1].\n  A plain-http request would send the key and the state unencrypted. Put an\n  https:// front on the deployment, or reach it through a port-forward to\n  localhost.\nXOR answered with a redirect — the endpoint replied 301, 302, 307 or 308.\n  The provider does not follow it and does not retry. Set the base URL to the\n  address the redirect leads to.\nThe XOR base URL is not a valid absolute URL — the value cannot be\n  parsed as a URL. Set it to an absolute origin.\nXOR requires an API key — set XOR_API_KEY, or pass\n  credentials.xor.apiKey.\nA 401 — the endpoint rejected the key. The provider instance stops\n  retrying after a rejection, so fix the key and construct a new one.\nA 403 with team_model_access_denied — the LiteLLM proxy's team for this\n  key does not allow the model you asked for (XOR_MODEL, or xor-1.1 by\n  default). Add the model to the team; the instance works again without a\n  restart.\nmax_tokens_exceeded — the state is over about 200,000 estimated tokens,\n  or the state, the questions and the media together are more than the\n  deployment's prefill. Shorten the state, or send fewer or smaller images or a\n  shorter video.","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"Troubleshooting","lvl3":""}},
-{"objectID":"cdd6e7c83dfa23afa5cf0bb7d23081881600f72a5ac64ca56d383736c209f285","title":"See also","url":"/docs/getting-started/providers/xor#see-also","content":"TypeSafe (Jev) Provider Guide\nLaya Provider Guide\nPerplexity Decisions Provider Guide\nCloudflare Clef Provider Guide\nSubQ Provider Guide\nOpenAI Decisions Provider Guide\nThe decide inference type\nXOR on Hugging Face","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"See also","lvl3":""}},
+{"objectID":"cdd6e7c83dfa23afa5cf0bb7d23081881600f72a5ac64ca56d383736c209f285","title":"See also","url":"/docs/getting-started/providers/xor#see-also","content":"TypeSafe (Jev) Provider Guide\nLaya Provider Guide\nPerplexity Decisions Provider Guide\nCloudflare Clef Provider Guide\nSubQ Provider Guide\nOpenAI Decisions Provider Guide\nScaleDown Provider Guide\nThe decide inference type\nXOR on Hugging Face","hierarchy":{"lvl0":"Getting Started","lvl1":"XOR Provider Guide","lvl2":"See also","lvl3":""}},
 {"objectID":"b62066defb7f4f21e315563cc367d3573fb4aacd3fa38685f76f1a3b110c0437","title":"Z.AI Provider Guide","url":"/docs/getting-started/providers/z-ai","content":"Z.AI Provider Guide\n\nZ.AI is a Tier-2 catalog provider: OpenAI-wire-compatible, so its entire\nintegration is one JSON file (src/lib/providers/catalog/z-ai.json) rather\nthan hand-written code. That file is the source of truth for everything on this\npage.\n\nVerification status: this entry is docs-verified only, not yet\nlive-verified. The vendor's GET https://api.z.ai/api/paas/v4/models needs\na key (it answers HTTP 401 without one), so the model ids come from Z.AI's\npublic docs and the roster has not been checked. No account was created and no\nAPI key was used to build this entry. evidence.liveMatrix is null until\nsomeone runs the live capability matrix with a real key (see\nVerification status below).\n\nKey Facts\nProvider id: z-ai (alias zai)\nProtocol: OpenAI-compatible (/chat/completions). The\n  OpenAI Python SDK page\n  says \"Z.AI provides interfaces compatible with OpenAI API\" and adds \"In some\n  scenarios, there are still differences between Z.AI and OpenAI interfaces, but\n  this does not affect overall compatibility.\"\nBase URL: https://api.z.ai/api/paas/v4 — the\n  Introduction page calls it\n  \"Z.ai Platform's general API endpoint\". Its warning says GLM Coding Plan users\n  follow the Coding Plan tutorial to configure a dedicated endpoint, and the\n  Coding Plan Endpoint Guide lists\n  https://api.z.ai/api/coding/paas/v4 for OpenAI Chat Completions. This entry\n  uses the general endpoint; Z_AI_BASE_URL overrides it.\nDefault model: glm-5.3 — the default of the model parameter in the\n  Chat Completion reference\n  and the model in its request examples\nModels in catalog: 12, curated from the 21 ids the Chat Completion\n  reference lists (14 in its text-model enum, 7 in its vision-model enum). The\n  nine ids left out to stay inside the 12-model cap are glm-4.5-x,\n  glm-4.5-airx, glm-4.5-flash, glm-4-32b-0414-128k, glm-4.6v,\n  glm-4.6v-flash, glm-4.6v-flashx, glm-4.5v and\n  autoglm-phone-multilingual\nStreaming: supported — the\n  Streaming Messages page\n  documents stream=True with Server-Sent Events, and the Chat Completion\n  reference says the Event Stream ends with a data: [DONE] message\nTool calling: model-dependent — the\n  Function Calling page\n  documents tools and a returned tool_calls array, and the Chat Completion\n  reference says of tool_choice \"The default value is auto, and only auto is\n  supported.\" For vision models it says tools are \"Only supported by\n  GLM-5.3-Flash series, the GLM-4.6V series, and autoglm-phone-multilingual\"\nTools while streaming: supported (true) — the\n  Thinking Mode page\n  gives an Interleaved Thinking + Tool Calling example that sends tools with\n  stream=True and reads tool_calls from the streamed deltas. The\n  Tool Streaming Output page\n  documents a separate tool_stream parameter, which this entry does not set.\n  The Thinking Mode page also says \"thinking blocks should be explicitly\n  preserved and returned together with the tool results.\", and its example\n  appends an assistant message carrying reasoning_content and tool_calls\n  before the tool message; this entry does not set\n  quirks.replayReasoningContent\nStructured output: supported — the\n  Structured Output page\n  documents response_format set to {\"type\": \"json_object\"}. The Chat\n  Completion reference lists text and json_object as the response_format\n  type values, so the entry sets quirks.responseFormatDowngrade to send\n  json_object for schema requests; generate({ schema }) still validates the\n  result client-side. No request with json_schema was sent. The Chat\n  Completion reference says of response_format \"Only text models support this\n  field.\", and the\n  GLM-5.3-Flash/FlashX page\n  lists Structured Output among its Capabilities (\"Supports structured output\n  formats such as JSON for seamless system integration.\")\nStructured output + tools together: not declared (false) — no combined\n  probe was possible without credentials\nVision: glm-5.3-flash and glm-5.3-flashx — the\n  GLM-5.3-Flash/FlashX page\n  shows Input Modality \"Video / Image / Text / File\". The GLM-5.3 page says\n  \"text-only inputs\", and the pages for the other nine catalog models show\n  Input Modalities \"Text\"\nEmbeddings: not declared on this catalog entry\nThinking: not declared (false) — the entry sets none of the thinking\n  or reasoning_effort request parameters. The Chat Completion reference\n  documents both, and the\n  GLM-5.3 page says \"GLM-5.3 always\n  operates with reasoning enabled\"\nBilling: schema value free-tier, see Billing below\nKey format: none declared. The Introduction page shows the header\n  Authorization: Bearer ZAI_API_KEY\n\nQuick Start\nGet an API key\nVisit: https://z.ai/manage-apikey/apikey-list — the Quick Start says to access the Z.AI Open Platform (https://z.ai/model-api) and \"Register or Login.\", then to \"Create an API Key\" in the API Keys management page and \"Copy your API Key for use.\"\nAuthentication is HTTP Bearer: the Introduction page (https://docs.z.ai/api-reference/introduction.md) shows the header Authorization: Bearer ZAIAPIKEY\n","hierarchy":{"lvl0":"Getting Started","lvl1":"Z.AI Provider Guide","lvl2":"","lvl3":""}},
 {"objectID":"ca9f10745365c71403f5bcba388684ac70aeb8f48c3d7b3fac85b832a4eead37","title":"Z.AI Provider Guide","url":"/docs/getting-started/providers/z-ai#zai-provider-guide","content":"Z.AI is a Tier-2 catalog provider: OpenAI-wire-compatible, so its entire\nintegration is one JSON file (src/lib/providers/catalog/z-ai.json) rather\nthan hand-written code. That file is the source of truth for everything on this\npage.\n\nVerification status: this entry is docs-verified only, not yet\nlive-verified. The vendor's GET https://api.z.ai/api/paas/v4/models needs\na key (it answers HTTP 401 without one), so the model ids come from Z.AI's\npublic docs and the roster has not been checked. No account was created and no\nAPI key was used to build this entry. evidence.liveMatrix is null until\nsomeone runs the live capability matrix with a real key (see\nVerification status below).","hierarchy":{"lvl0":"Getting Started","lvl1":"Z.AI Provider Guide","lvl2":"Z.AI Provider Guide","lvl3":""}},
 {"objectID":"d509ecad3c62c73f008d7e2f88de10168c19308f90c63101973f3ae4983a7318","title":"Key Facts","url":"/docs/getting-started/providers/z-ai#key-facts","content":"Provider id: z-ai (alias zai)\nProtocol: OpenAI-compatible (/chat/completions). The\n  OpenAI Python SDK page\n  says \"Z.AI provides interfaces compatible with OpenAI API\" and adds \"In some\n  scenarios, there are still differences between Z.AI and OpenAI interfaces, but\n  this does not affect overall compatibility.\"\nBase URL: https://api.z.ai/api/paas/v4 — the\n  Introduction page calls it\n  \"Z.ai Platform's general API endpoint\". Its warning says GLM Coding Plan users\n  follow the Coding Plan tutorial to configure a dedicated endpoint, and the\n  Coding Plan Endpoint Guide lists\n  https://api.z.ai/api/coding/paas/v4 for OpenAI Chat Completions. This entry\n  uses the general endpoint; Z_AI_BASE_URL overrides it.\nDefault model: glm-5.3 — the default of the model parameter in the\n  Chat Completion reference\n  and the model in its request examples\nModels in catalog: 12, curated from the 21 ids the Chat Completion\n  reference lists (14 in its text-model enum, 7 in its vision-model enum). The\n  nine ids left out to stay inside the 12-model cap are glm-4.5-x,\n  glm-4.5-airx, glm-4.5-flash, glm-4-32b-0414-128k, glm-4.6v,\n  glm-4.6v-flash, glm-4.6v-flashx, glm-4.5v and\n  autoglm-phone-multilingual\nStreaming: supported — the\n  Streaming Messages page\n  documents stream=True with Server-Sent Events, and the Chat Completion\n  reference says the Event Stream ends with a data: [DONE] message\nTool calling: model-dependent — the\n  Function Calling page\n  documents tools and a returned tool_calls array, and the Chat Completion\n  reference says of tool_choice \"The default value is auto, and only auto is\n  supported.\" For vision models it says tools are \"Only supported by\n  GLM-5.3-Flash series, the GLM-4.6V series, and autoglm-phone-multilingual\"\nTools while streaming: supported (true) — the\n  Thinking Mode page\n  gives an Interleaved Thinking + Tool Calling example that sends tools with\n  stream=True and reads tool_calls from the streamed deltas.","hierarchy":{"lvl0":"Getting Started","lvl1":"Z.AI Provider Guide","lvl2":"Key Facts","lvl3":""}},
@@ -8725,7 +8753,7 @@
 {"objectID":"95f574ed3e7979eee5be188ccd16efb2639e30e8258e55366d7ccd344bbc3126","title":"Related Documentation","url":"/docs/implementation-guides/14-rag-document-processing#related-documentation","content":"Vector Store Integrations\nEvaluation and Scoring\nMaster Implementation Guide","hierarchy":{"lvl0":"Implementation Guides","lvl1":"RAG Document Processing - Implementation Guide","lvl2":"Related Documentation","lvl3":""}},
 {"objectID":"60877eba17c1fe5c9fda2100a737f42fddb5c4c7083297e61a66884dbe5c326e","title":"NeuroLink","url":"/docs/","content":"🧠 NeuroLink\n  The Pipe Layer of an AI Nervous System\n  Provider Neurons Across Major AI Vendors | 3 Inference Types (generate · stream · decide) | Voice (TTS/STT/Realtime) | 58+ MCP Servers | HITL Security | Redis Persistence\n\nNeuroLink is the pipe layer of an AI nervous system: one interface connecting provider neurons — major AI vendors and local runtimes — to the applications that consume them. Built-in tooling and an opinionated factory architecture mean adding a new provider, or a new capability, never touches application code. NeuroLink ships as both a TypeScript SDK and a professional CLI so teams can build, operate, and iterate on AI features quickly.\n\n🧠 What is NeuroLink?\n\nNeuroLink is the pipe layer of an AI nervous system. Providers — OpenAI, Anthropic, Google, AWS, Azure, DeepSeek, NVIDIA NIM, local runtimes like Ollama and llama.cpp, and dozens more — are the neurons: each generates a different kind of intelligence, at a different cost and latency. NeuroLink is the vascular layer that carries that intelligence, as a stream, to the applications that consume it, across three inference types: generate and stream produce text, decide produces a calibrated boolean/choice/score judgment instead.\n\nExtracted from production systems at Juspay, NeuroLink provides a practical, TypeScript-first way to plug any application into that nervous system. Switch which neuron answers a request with a single parameter change — any provider you're building with, or any provider you add.\n\nWhy NeuroLink? Three genuine inference types, not one dressed up three ways — generate and stream produce text, while decide returns a typed, calibrated judgment (boolean / choice / score) with no text at all, for the routing and gating decisions the other two were never meant to make. Every neuron plugs into the same pipe. Switch providers with a single parameter change, leverage 64+ built-in tools and MCP servers, deploy with confidence using enterprise features like Redis memory and multi-provider failover, and optimize costs automatically with intelligent routing. Use it via our professional CLI or TypeScript SDK—whichever fits your workflow.\n\nWhere we're headed: We're building for the future of AI—edge-first execution and continuous streaming architectures that make AI practically free and universally available. Read our vision →\n\nGet Started in \\ Observability Guide\nServer Adapters -- Deploy NeuroLink as an HTTP API server with your framework of choice (Hono, Express, Fastify, Koa). Full CLI support with serve and server commands for foreground/background modes, route management, and OpenAPI generation. -> Server Adapters Guide\nTitle Generation Events -- Emit real-time events when conversation titles are auto-generated. Listen to conversation:titleGenerated for session tracking. -> Conversation Memory Guide\nCustom Title Prompts -- Customize conversation title generation with NEUROLINK_TITLE_PROMPT environment variable. Use ${userMessage} placeholder for dynamic prompts. -> Conversation Memory Guide\nVideo Generation -- Transform images into 8-second videos with synchronized audio using Google Veo 3.1 via Vertex AI. Supports 720p/1080p resolutions, portrait/landscape aspect ratios. -> Video Generation Guide\nImage Generation -- Generate images from text prompts using Gemini models via Vertex AI or Google AI Studio. Supports streaming mode with automatic file saving. -> Image Generation Guide\nHTTP/Streamable HTTP Transport for MCP -- Connect to remote MCP servers via HTTP with authentication headers, retry logic, and rate limiting. -> HTTP Transport Guide\nClaude Subscription (OAuth) Support -- Use your Claude Pro/Max/Team subscription with NeuroLink via OAuth authentication, no API key required. -> Subscription Guide\nGemini 3 Preview Support - Full support for gemini-3-flash-preview and gemini-3-pro-preview with extended thinking capabilities\nStructured Output with Zod Schemas -- Type-safe JSON generation with automatic validation using schema + output.format: \"json\" in generate(). -> Structured Output Guide\nCSV File Support -- Attach CSV files to prompts for AI-powered data analysis with auto-detection. -> CSV Guide\nPDF File Support -- Process PDF documents with native visual analysis for Vertex AI, Anthropic, Bedrock, AI Studio. -> PDF Guide\n50+ File Types -- Process Excel, Word, RTF, JSON, YAML, XML, HTML, SVG, Markdown, and 50+ code languages with intelligent content extraction. -> File Processors Guide\nLiteLLM Integration -- Access 100+ AI models across a broad range of AI providers through unified interface. -> Setup Guide\nSageMaker Integration -- Deploy and use custom trained models on AWS infrastructure. -> Setup Guide\nOpenRouter Integration -- Access 300+ models from OpenAI, Anthropic, Google, Meta, and more through a single unified API. -> Setup Guide\nHuman-in-the-loop workflows -- Pause generation for user approval/input before tool execution. -> HITL Guide\nGuardrails middleware -- Block PII, profanity, and unsafe cont","hierarchy":{"lvl0":"Docs","lvl1":"NeuroLink","lvl2":"","lvl3":""}},
 {"objectID":"7bd402837d03dbfbb9586674bc914cfc6ee7691e33540eeabe6555363697375e","title":"🧠 What is NeuroLink?","url":"/docs/#-what-is-neurolink","content":"NeuroLink is the pipe layer of an AI nervous system. Providers — OpenAI, Anthropic, Google, AWS, Azure, DeepSeek, NVIDIA NIM, local runtimes like Ollama and llama.cpp, and dozens more — are the neurons: each generates a different kind of intelligence, at a different cost and latency. NeuroLink is the vascular layer that carries that intelligence, as a stream, to the applications that consume it, across three inference types: generate and stream produce text, decide produces a calibrated boolean/choice/score judgment instead.\n\nExtracted from production systems at Juspay, NeuroLink provides a practical, TypeScript-first way to plug any application into that nervous system. Switch which neuron answers a request with a single parameter change — any provider you're building with, or any provider you add.\n\nWhy NeuroLink? Three genuine inference types, not one dressed up three ways — generate and stream produce text, while decide returns a typed, calibrated judgment (boolean / choice / score) with no text at all, for the routing and gating decisions the other two were never meant to make. Every neuron plugs into the same pipe. Switch providers with a single parameter change, leverage 64+ built-in tools and MCP servers, deploy with confidence using enterprise features like Redis memory and multi-provider failover, and optimize costs automatically with intelligent routing. Use it via our professional CLI or TypeScript SDK—whichever fits your workflow.\n\nWhere we're headed: We're building for the future of AI—edge-first execution and continuous streaming architectures that make AI practically free and universally available. Read our vision →\n\nGet Started in \\<5 Minutes →","hierarchy":{"lvl0":"Docs","lvl1":"NeuroLink","lvl2":"🧠 What is NeuroLink?","lvl3":""}},
-{"objectID":"293082e7b04b9b0eb186f10fec40eefc1996d8eab9becff9e2b428731642291f","title":"What's New (Q1 2026)","url":"/docs/#whats-new-q1-2026","content":"| Feature                            | Version | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     | Guide                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |\n| ---------------------------------- | ------- |","hierarchy":{"lvl0":"Docs","lvl1":"NeuroLink","lvl2":"What's New (Q1 2026)","lvl3":""}},
+{"objectID":"293082e7b04b9b0eb186f10fec40eefc1996d8eab9becff9e2b428731642291f","title":"What's New (Q1 2026)","url":"/docs/#whats-new-q1-2026","content":"| Feature                            | Version | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             | Guide                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |\n| ---------------------------------- | ------- |","hierarchy":{"lvl0":"Docs","lvl1":"NeuroLink","lvl2":"What's New (Q1 2026)","lvl3":""}},
 {"objectID":"3e5ecc512d348470d92717ef33a2ff5067ffdd99c7ea212613df7986db133390","title":"Enterprise Security: Human-in-the-Loop (HITL)","url":"/docs/#enterprise-security-human-in-the-loop-hitl","content":"NeuroLink includes a HITL (Human-in-the-Loop) system for regulated industries and high-stakes AI operations:\n\n| Capability                  | Description                                                             | Use Case                                   |\n| --------------------------- | ----------------------------------------------------------------------- | ------------------------------------------ |\n| Tool Approval Workflows | Require human approval before AI executes sensitive tools               | Financial transactions, data modifications |\n| Output Validation       | Route AI outputs through human review pipelines                         | Medical diagnosis, legal documents         |\n| Confidence Thresholds   | Automatically trigger human review below confidence level               | Critical business decisions                |\n| Complete Audit Trail    | Audit logging to support your compliance program (HIPAA / SOC 2 / GDPR) | Regulated industries                       |\n\nEnterprise HITL Guide | Quick Start","hierarchy":{"lvl0":"Docs","lvl1":"NeuroLink","lvl2":"Enterprise Security: Human-in-the-Loop (HITL)","lvl3":""}},
 {"objectID":"c062d596f0f8940391d038563717712409b6055fdf70d92ab51cb35caf672663","title":"Get Started in Two Steps","url":"/docs/#get-started-in-two-steps","content":"`bash","hierarchy":{"lvl0":"Docs","lvl1":"NeuroLink","lvl2":"Get Started in Two Steps","lvl3":""}},
 {"objectID":"e9e7167316a23ea5fd5db23894b2e52405ba61f5ab57cc986d155490dd325674","title":"1. Run the interactive setup wizard (select providers, validate keys)","url":"/docs/#1-run-the-interactive-setup-wizard-select-providers-validate-keys","content":"pnpm dlx @juspay/neurolink setup","hierarchy":{"lvl0":"Docs","lvl1":"NeuroLink","lvl2":"1. Run the interactive setup wizard (select providers, validate keys)","lvl3":""}},
@@ -9839,11 +9867,11 @@
 {"objectID":"205e4e44d27692634aa5454565d9f060ceb88cdac97e7d37622bc5b9a1eefeb6","title":"Verification commands","url":"/docs/provider-integration/tiers/tier-1-aggregator-passthrough#verification-commands","content":"`bash","hierarchy":{"lvl0":"Provider Integration","lvl1":"Tier 1 — Aggregator Passthrough","lvl2":"Verification commands","lvl3":""}},
 {"objectID":"09ead22698863d776e82e2f0f3dfdd9a596f39cc106302a02c51a6ef37b436de","title":"LiteLLM example","url":"/docs/provider-integration/tiers/tier-1-aggregator-passthrough#litellm-example","content":"pnpm run cli generate \"hello\" --provider litellm --model","hierarchy":{"lvl0":"Provider Integration","lvl1":"Tier 1 — Aggregator Passthrough","lvl2":"LiteLLM example","lvl3":""}},
 {"objectID":"f46ff1668fbf30431a2b6578654212feb2043845eb7a28281e47075703948472","title":"OpenRouter example","url":"/docs/provider-integration/tiers/tier-1-aggregator-passthrough#openrouter-example","content":"pnpm run cli generate \"hello\" --provider openrouter --model /\n\n\nBoth should return a normal GenerateResult with non-empty content`. If\neither 400s with an \"unknown model\" style error, the aggregator doesn't\nactually serve that model yet — fix the aggregator-side config, not\nNeuroLink.","hierarchy":{"lvl0":"Provider Integration","lvl1":"Tier 1 — Aggregator Passthrough","lvl2":"OpenRouter example","lvl3":""}},
-{"objectID":"0c443abf0cfa1b31edc3337559bddbf37453d87732cfd17b08d461f55c77b195","title":"Tier 2 — Catalog Entry","url":"/docs/provider-integration/tiers/tier-2-catalog-entry","content":"Tier 2 — Catalog Entry\n\nWhen this applies: the vendor speaks the OpenAI /v1/chat/completions\nwire format (Bearer auth, standard SSE, standard JSON body) and needs\nzero behavioral overrides — no custom adjustRequestBody, no\nadjustResponseFormat, no nonstandard auth header, no adjustBodyAfter400.\nThis is the Groq/xAI/Together AI/Fireworks/Perplexity/Cloudflare/Mistral\nshape from before the redesign — now expressed as one JSON file instead of\na hand-written subclass (see ../adr/0002-catalog-over-subclass-default.md\nand the single-JSON spec at\ndocs/superpowers/plans/2026-08-28-provider-json-catalog-spec.md).\n\nIf you're not sure whether your provider is quirk-free, start writing the\nJSON anyway — if it turns out you need a hook, migrate to Tier 3 instead of\nforcing the quirk into the catalog shape.\n\nFiles touched (end state)\n\n| #   | File                                  | Change                          |\n| --- | ------------------------------------- | ------------------------------- |\n| 1   | src/lib/providers/catalog/<id>.json | The entire integration, as data |\n\nThat is the whole list. Everything else is generated or derived:\npnpm run codegen:catalog writes the AIProviderName member, the\n  <Name>Models enum and the NeurolinkCredentials key into marked\n  regions, plus the generated catalog index. Pre-commit and CI fail on\n  stale output, so it can't drift.\nThe runtime loader builds the registry entry, descriptor, config options,\n  context windows, pricing, vision map and model choices from the JSON.\nThe test suites derive their spec rows, matrix rows and counts from the\n  built catalog.\n\nScaffold it with:\n\nTier 2 emits exactly two files: a pre-filled <id>.json (with TODO markers\nwhere only a live probe can supply the truth) and a short checklist. The\nJSON deliberately fails schema validation until every TODO is replaced.\n\nCount pins — nothing to bump\n\nThere are none left. The wiring and descriptor suites compute their\nexpected totals from CATALOG_PROVIDER_IDS, and the matrix rows spread\nfrom the catalog with a completeness guard that throws if any catalog\nprovider is missing. Adding a provider changes no test file.\n\nsrc/lib/factories/providerRegistry.ts isn't in the table either:\n_doRegister() loops over the catalog and registers every entry\ngenerically.\n\nOne historical exception worth re-checking: commandFactory.ts's separate\nsetup [provider] subcommand hand-hardcoded its own provider-choices array.\nThat was migrated to derive from the enum (PR #1583) — confirm it still\ndoes before assuming you need a manual edit.\n\nThe JSON, field by field\n\nid, displayName, aliases, tier, then:\n\nwire — baseURL for the normal case. Env var names derive by\nconvention (<CONSTANT_CASE(id)>_API_KEY / _BASE_URL / _MODEL); set\nenvOverrides only for a vendor that breaks it. For a URL computed from\nanother credential (Cloudflare's account id) use baseURLTemplate plus\nexactly one extraCredentials entry.\n\nmodels — default, fallbacks, defaultContextWindow,\ndefaultMaxOutputTokens, and a catalog map of model id → {contextWindow?,\nmaxOutputTokens?, pricingPerMTok?, vision, tools?, status, description, enumMember?}.\nOmit a number rather than invent one. Optional refinements:\n\n| Field                  | Use it when                                                                                                                                            |\n| ---------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------ |\n| enumMember           | The derived constant-case name would break an existing export                                                                                          |\n| enumTypeName         | The <Name>Models enum name must differ from the derived one                                                                                          |\n| fallbackModelName    | The legacy fallback differs from fallbacks[1] ?? fallbacks[0]                                                                                        |\n| registryDefaultModel | The registry default differs from default                                                                                                            |\n| topModels            | The CLI picker should show a curated ordered subset                                                                                                    |\n| visionModel          | Vision tests need a specific model (the default is text-only)                                                                                          |\n| testModel            | The catalog default is retired/gated on the testing account                                                                                            |\n| tools (per model)    | capabilities.tools is \"model-dependent\" and the vendor's own docs say which served models actually accept tool definitions (Reka: flash-3 doesn't) |\n\ntestModel is matrix-only","hierarchy":{"lvl0":"Provider Integration","lvl1":"Tier 2 — Catalog Entry","lvl2":"","lvl3":""}},
+{"objectID":"0c443abf0cfa1b31edc3337559bddbf37453d87732cfd17b08d461f55c77b195","title":"Tier 2 — Catalog Entry","url":"/docs/provider-integration/tiers/tier-2-catalog-entry","content":"Tier 2 — Catalog Entry\n\nWhen this applies: the vendor speaks the OpenAI /v1/chat/completions\nwire format (Bearer auth, standard SSE, standard JSON body) and needs\nzero behavioral overrides — no custom adjustRequestBody, no\nadjustResponseFormat, no nonstandard auth header, no adjustBodyAfter400.\nThis is the Groq/xAI/Together AI/Fireworks/Perplexity/Cloudflare/Mistral\nshape from before the redesign — now expressed as one JSON file instead of\na hand-written subclass (see ../adr/0002-catalog-over-subclass-default.md\nand the single-JSON spec at\ndocs/superpowers/plans/2026-08-28-provider-json-catalog-spec.md).\n\nIf you're not sure whether your provider is quirk-free, start writing the\nJSON anyway — if it turns out you need a hook, migrate to Tier 3 instead of\nforcing the quirk into the catalog shape.\n\nFiles touched (end state)\n\n| #   | File                                  | Change                          |\n| --- | ------------------------------------- | ------------------------------- |\n| 1   | src/lib/providers/catalog/<id>.json | The entire integration, as data |\n\nThat is the whole list. Everything else is generated or derived:\npnpm run codegen:catalog writes the AIProviderName member, the\n  <Name>Models enum and the NeurolinkCredentials key into marked\n  regions, plus the generated catalog index. Pre-commit and CI fail on\n  stale output, so it can't drift.\nThe runtime loader builds the registry entry, descriptor, config options,\n  context windows, pricing, vision map and model choices from the JSON.\nThe test suites derive their spec rows, matrix rows and counts from the\n  built catalog.\n\nScaffold it with:\n\nTier 2 emits exactly two files: a pre-filled <id>.json (with TODO markers\nwhere only a live probe can supply the truth) and a short checklist. The\nJSON deliberately fails schema validation until every TODO is replaced.\n\nCount pins — nothing to bump\n\nThere are none left. The wiring and descriptor suites compute their\nexpected totals from CATALOG_PROVIDER_IDS, and the matrix rows spread\nfrom the catalog with a completeness guard that throws if any catalog\nprovider is missing. Adding a provider changes no test file.\n\nsrc/lib/factories/providerRegistry.ts isn't in the table either:\n_doRegister() loops over the catalog and registers every entry\ngenerically.\n\nOne historical exception worth re-checking: commandFactory.ts's separate\nsetup [provider] subcommand hand-hardcoded its own provider-choices array.\nThat was migrated to derive from the enum (PR #1583) — confirm it still\ndoes before assuming you need a manual edit.\n\nThe JSON, field by field\n\nid, displayName, aliases, tier, then:\n\nwire — baseURL for the normal case. Env var names derive by\nconvention (<CONSTANT_CASE(id)>_API_KEY / _BASE_URL / _MODEL); set\nenvOverrides only for a vendor that breaks it. For a URL computed from\nanother credential (Cloudflare's account id) use baseURLTemplate plus\nexactly one extraCredentials entry.\n\nmodels — default, fallbacks, defaultContextWindow,\ndefaultMaxOutputTokens, and a catalog map of model id → {contextWindow?,\nmaxOutputTokens?, pricingPerMTok?, vision, tools?, status, description, enumMember?}.\nOmit a number rather than invent one. A per-model maxOutputTokens is enforced,\nnot just displayed: when the provider has no PROVIDER_MAX_TOKENS row,\ngetSafeMaxTokens clamps a larger requested maxTokens to it, so record only a\nlimit the vendor states and leave it out when the figure looks derived from the\ncontext window. defaultMaxOutputTokens is not enforced this way. Optional\nrefinements:\n\n| Field                  | Use it when                                                                                                                                            |\n| ---------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------ |\n| enumMember           | The derived constant-case name would break an existing export                                                                                          |\n| enumTypeName         | The <Name>Models enum name must differ from the derived one                                                                                          |\n| fallbackModelName    | The legacy fallback differs from fallbacks[1] ?? fallbacks[0]                                                                                        |\n| registryDefaultModel | The registry default differs from default                                                                                                            |\n| topModels            | The CLI picker should show a curated ordered subset                                                                                                    |\n| visionModel          | Vision tests need a specific model (the default is text-only)                                                                                          |\n| testModel            | The catalog default","hierarchy":{"lvl0":"Provider Integration","lvl1":"Tier 2 — Catalog Entry","lvl2":"","lvl3":""}},
 {"objectID":"4b278bfa53040dfc60deaffd7c0042488d914cc4eff89ded532f5582eceadce6","title":"Tier 2 — Catalog Entry","url":"/docs/provider-integration/tiers/tier-2-catalog-entry#tier-2-catalog-entry","content":"When this applies: the vendor speaks the OpenAI /v1/chat/completions\nwire format (Bearer auth, standard SSE, standard JSON body) and needs\nzero behavioral overrides — no custom adjustRequestBody, no\nadjustResponseFormat, no nonstandard auth header, no adjustBodyAfter400.\nThis is the Groq/xAI/Together AI/Fireworks/Perplexity/Cloudflare/Mistral\nshape from before the redesign — now expressed as one JSON file instead of\na hand-written subclass (see ../adr/0002-catalog-over-subclass-default.md\nand the single-JSON spec at\ndocs/superpowers/plans/2026-08-28-provider-json-catalog-spec.md).\n\nIf you're not sure whether your provider is quirk-free, start writing the\nJSON anyway — if it turns out you need a hook, migrate to Tier 3 instead of\nforcing the quirk into the catalog shape.","hierarchy":{"lvl0":"Provider Integration","lvl1":"Tier 2 — Catalog Entry","lvl2":"Tier 2 — Catalog Entry","lvl3":""}},
 {"objectID":"1bb614eb5848cace52afb691aa99aeab78366a84a216f758ef9df6c86e49c0cc","title":"Files touched (end state)","url":"/docs/provider-integration/tiers/tier-2-catalog-entry#files-touched-end-state","content":"| #   | File                                  | Change                          |\n| --- | ------------------------------------- | ------------------------------- |\n| 1   | src/lib/providers/catalog/<id>.json | The entire integration, as data |\n\nThat is the whole list. Everything else is generated or derived:\npnpm run codegen:catalog writes the AIProviderName member, the\n  <Name>Models enum and the NeurolinkCredentials key into marked\n  regions, plus the generated catalog index. Pre-commit and CI fail on\n  stale output, so it can't drift.\nThe runtime loader builds the registry entry, descriptor, config options,\n  context windows, pricing, vision map and model choices from the JSON.\nThe test suites derive their spec rows, matrix rows and counts from the\n  built catalog.\n\nScaffold it with:\n\nTier 2 emits exactly two files: a pre-filled <id>.json (with TODO markers\nwhere only a live probe can supply the truth) and a short checklist. The\nJSON deliberately fails schema validation until every TODO is replaced.","hierarchy":{"lvl0":"Provider Integration","lvl1":"Tier 2 — Catalog Entry","lvl2":"Files touched (end state)","lvl3":""}},
 {"objectID":"9627a6b63ab505017fbc3edd7f9ec312f548b35aef74edd6e2be5eac8e1ade0f","title":"Count pins — nothing to bump","url":"/docs/provider-integration/tiers/tier-2-catalog-entry#count-pins-nothing-to-bump","content":"There are none left. The wiring and descriptor suites compute their\nexpected totals from CATALOG_PROVIDER_IDS, and the matrix rows spread\nfrom the catalog with a completeness guard that throws if any catalog\nprovider is missing. Adding a provider changes no test file.\n\nsrc/lib/factories/providerRegistry.ts isn't in the table either:\n_doRegister() loops over the catalog and registers every entry\ngenerically.\n\nOne historical exception worth re-checking: commandFactory.ts's separate\nsetup [provider] subcommand hand-hardcoded its own provider-choices array.\nThat was migrated to derive from the enum (PR #1583) — confirm it still\ndoes before assuming you need a manual edit.","hierarchy":{"lvl0":"Provider Integration","lvl1":"Tier 2 — Catalog Entry","lvl2":"Count pins — nothing to bump","lvl3":""}},
-{"objectID":"672b4a181677219d0db0d64d290a3f1a268691c336f02fe776a01d94d0684ae5","title":"The JSON, field by field","url":"/docs/provider-integration/tiers/tier-2-catalog-entry#the-json-field-by-field","content":"id, displayName, aliases, tier, then:\n\nwire — baseURL for the normal case. Env var names derive by\nconvention (<CONSTANT_CASE(id)>_API_KEY / _BASE_URL / _MODEL); set\nenvOverrides only for a vendor that breaks it. For a URL computed from\nanother credential (Cloudflare's account id) use baseURLTemplate plus\nexactly one extraCredentials entry.\n\nmodels — default, fallbacks, defaultContextWindow,\ndefaultMaxOutputTokens, and a catalog map of model id → {contextWindow?,\nmaxOutputTokens?, pricingPerMTok?, vision, tools?, status, description, enumMember?}.\nOmit a number rather than invent one. Optional refinements:\n\n| Field                  | Use it when                                                                                                                                            |\n| ---------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------ |\n| enumMember           | The derived constant-case name would break an existing export                                                                                          |\n| enumTypeName         | The <Name>Models enum name must differ from the derived one                                                                                          |\n| fallbackModelName    | The legacy fallback differs from fallbacks[1] ??","hierarchy":{"lvl0":"Provider Integration","lvl1":"Tier 2 — Catalog Entry","lvl2":"The JSON, field by field","lvl3":""}},
+{"objectID":"672b4a181677219d0db0d64d290a3f1a268691c336f02fe776a01d94d0684ae5","title":"The JSON, field by field","url":"/docs/provider-integration/tiers/tier-2-catalog-entry#the-json-field-by-field","content":"id, displayName, aliases, tier, then:\n\nwire — baseURL for the normal case. Env var names derive by\nconvention (<CONSTANT_CASE(id)>_API_KEY / _BASE_URL / _MODEL); set\nenvOverrides only for a vendor that breaks it. For a URL computed from\nanother credential (Cloudflare's account id) use baseURLTemplate plus\nexactly one extraCredentials entry.\n\nmodels — default, fallbacks, defaultContextWindow,\ndefaultMaxOutputTokens, and a catalog map of model id → {contextWindow?,\nmaxOutputTokens?, pricingPerMTok?, vision, tools?, status, description, enumMember?}.\nOmit a number rather than invent one. A per-model maxOutputTokens is enforced,\nnot just displayed: when the provider has no PROVIDER_MAX_TOKENS row,\ngetSafeMaxTokens clamps a larger requested maxTokens to it, so record only a\nlimit the vendor states and leave it out when the figure looks derived from the\ncontext window. defaultMaxOutputTokens is not enforced this way. Optional\nrefinements:\n\n| Field                  | Use it when                                                                                                                                            |\n| ---------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------ |\n| enumMember           | The derived constant-case name would break an existing export                                                                                          |\n| enumTypeName         | The <Name>Models enum name must differ from the derived one                                                                                          |\n| fallbackModelName    | The legacy fallback differs from fallbacks[1] ??","hierarchy":{"lvl0":"Provider Integration","lvl1":"Tier 2 — Catalog Entry","lvl2":"The JSON, field by field","lvl3":""}},
 {"objectID":"5f38f36d8546533723645d1643ea69418d7e5cbe47cde5d18b5ad998d053e476","title":"Live verification","url":"/docs/provider-integration/tiers/tier-2-catalog-entry#live-verification","content":"The mocked gates prove the wire contract, not the commercial reality.\nEach of these caught a real defect on the cerebras pilot:\nRoster probe first — authenticated GET <baseURL>/models. Pick\n   default, fallbacks and catalog keys from what the API serves TODAY.\n   Vendor docs listed four cerebras models; the live roster had two, and\n   the documented default 404'd (finding #7). This ages: both Groq's and\n   Fireworks' catalog defaults went dead within weeks.\nBilling policy — confirm how a working key is obtained. Cerebras has\n   no keyless free tier: even the \"$5 free credits\" require saving a payment\n   card (finding #8); SambaNova requires payment outright (finding #11).\nCapability probes — probe tools + response_format in one request\n   before setting structuredOutputWithTools; strict backends 400. Don't\n   copy another provider's flags on vibes.\nLive matrix — with a working key:\n   npx tsx test/continuous-test-suite-provider-matrix.ts --provider=<id>\n   must pass generate, stream, tool calling and structured output, and a\n   bare node dist/cli/index.js generate \"...\" --provider <id> must resolve\n   the default model. The pilot's first live run was 2/4 and surfaced an\n   SDK-wide bug (tool_choice emitted on tools-less requests — fixed in\n   #1564, now pinned by the mocked suite), which is why this step exists.\n\nRecord the outcome in evidence.liveMatrix.","hierarchy":{"lvl0":"Provider Integration","lvl1":"Tier 2 — Catalog Entry","lvl2":"Live verification","lvl3":""}},
 {"objectID":"bea00545b1053aa393e34f487097636d2ee23cfdc652c6724dc6029192017ffd","title":"Verification commands","url":"/docs/provider-integration/tiers/tier-2-catalog-entry#verification-commands","content":"All must exit 0 before opening the PR. The first three test commands run in\nthe provider-safety-net CI job (.github/workflows/ci.yml) — zero-API,\nzero-credential checks, so there's no reason to skip them locally. Add a\ncase to test:error-classifier-contract first if your entry sets\ntimeoutErrorClass or vendor-specific errorRules.\n\nBecause the suites derive from data, a new provider needs no new assertions\n— but if you ever do add one, run the break-one-assertion ritual: flip it,\nconfirm the suite reports ✗ and exits non-zero (not ⊘ skip — see the\nassertion-message hazard in CLAUDE.md), then restore it.","hierarchy":{"lvl0":"Provider Integration","lvl1":"Tier 2 — Catalog Entry","lvl2":"Verification commands","lvl3":""}},
 {"objectID":"3f8e5102409b953258a82548fec14a7726774f6e1e5c3c7d826400ef24ebf105","title":"Tier 3 — Adapter-Based Native","url":"/docs/provider-integration/tiers/tier-3-adapter-native","content":"Tier 3 — Adapter-Based Native\n\nWhen this applies: the vendor has its own SDK or wire format that\nisn't OpenAI-compatible, but it's still a normal request/response (or\nrequest/SSE-stream) HTTP+JSON lifecycle you can drive from a provider\nclass. This is the Anthropic/Google AI Studio shape — a dedicated\nsrc/lib/providers/<name>.ts extending BaseProvider directly, not the\nOpenAIChatCompletionsProvider family.\n\nCorrected from the original plan text (2026-08-18): the plan's\ndraft cited \"the Mistral/Cohere/Ollama shape\" as the Tier 3 example.\nThat's stale — mistral.ts, cohere.ts, and ollama/client.ts all\nextend OpenAIChatCompletionsProvider, the Tier 2 family, not\nBaseProvider directly. (Mistral is separately already named in\n../adr/0002-catalog-over-subclass-default.md as a Tier 2\ncatalog-migration candidate — the original draft was internally\ninconsistent about which tier Mistral belongs to.) The verified,\ncurrently-shipping examples of a chat/text provider extending\nBaseProvider directly are Anthropic\n(src/lib/providers/anthropic/client.ts) and Google AI Studio\n(src/lib/providers/googleAiStudio/client.ts) — used below.\nWhy these two, specifically: they implement the request/response and\nstreaming lifecycle against their vendor's own wire format directly,\ninside the provider class itself — they don't inherit that lifecycle\nfrom OpenAIChatCompletionsProvider's shared chat-completions\nimplementation the way Mistral/Cohere/Ollama do. That's the actual line\nbetween Tier 2 and Tier 3: not \"does the vendor have a custom SDK\" but\n\"does this class implement the provider surface itself, or inherit it.\"\nWhen you compare your new provider's shape against Anthropic/Google AI\nStudio, that's the property you're matching — not their specific\nrequest/response format, which is vendor-idiosyncratic and won't look\nlike yours.\n\nAs of 2026-08-18 there is no shared streaming-loop adapter beyond\nBaseProvider and OpenAIChatCompletionsProvider (checked\nsrc/lib/providers/ for a nativeAdapter*.ts or similar — none exists).\nIf one lands later, extend it instead of hand-rolling the SSE parser and\nmulti-step tool loop; the steps below describe the always-true minimum\nregardless of whether that shared adapter exists yet.\n\nFiles touched (end state)\n\n| #   | File                                                                                           | Change                                                                                                                                                                                                                    |\n| --- | ---------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| 1   | src/lib/constants/enums.ts                                                                   | New AIProviderName member + a <Name>Models enum (default + fallback model ids — Tier 3 providers keep an explicit model catalog since there's no OPENAI_COMPAT_CATALOG row to hold defaultModel/fallbackModels) |\n| 2   | src/lib/providers/<name>.ts                                                                  | NEW provider class extending BaseProvider (or a shared adapter, if one has landed by the time you read this)                                                                                                            |\n| 3   | src/lib/factories/providerDescriptors.ts                                                     | One ProviderDescriptor entry                                                                                                                                                                                            |\n| 4   | src/lib/factories/providerRegistry.ts                                                        | One ProviderFactory.registerProvider() block, dynamic import, 5-argument call including the descriptor argument from PROVIDER_DESCRIPTORS_BY_NAME                                                                   |\n| 5   | src/lib/types/providers.ts                                                                   | New NeurolinkCredentials[\"<key>\"] slice                                                                                                                                                                                 |\n| 6   | src/lib/adapters/providerImageAdapter.ts                                                     | VISION_CAPABILITIES entry — only if the provider/model is multimodal                                                                                                                                                    |\n| 7   | test/continuous-test-suite-providers-mocked.ts                                               | Mocked-contract section                                                        ","hierarchy":{"lvl0":"Provider Integration","lvl1":"Tier 3 — Adapter-Based Native","lvl2":"","lvl3":""}},
@@ -10052,10 +10080,10 @@
 {"objectID":"4578b51cac195eb59c8ac256ef9f2491197666719997b358b74e2b57b5a414fb","title":"TTS / Realtime Errors","url":"/docs/reference/error-codes#tts-realtime-errors","content":"TTSError and RealtimeError carry provider-specific messages (e.g.\nsynthesis failure, WebSocket disconnect, function-call failure). Inspect\nerror.cause for the underlying provider error.","hierarchy":{"lvl0":"Reference","lvl1":"Error Code Reference","lvl2":"TTS / Realtime Errors","lvl3":""}},
 {"objectID":"3e6c6ffd80fa23ecf3686bacf0db97fb3e7e6f430e82b2523745c88e257b6c45","title":"Common Triggers","url":"/docs/reference/error-codes#common-triggers","content":"STT_INVALID_AUDIO_FORMAT is thrown by STTProcessor.transcribe()\n  when options.format doesn't appear in the provider's\n  getSupportedFormats() list. The CLI infers the format from the\n  --input-audio file extension; the SDK requires you to pass it explicitly.\n  Fix: either convert the audio to a supported format, or use a different\n  STT provider. See docs/getting-started/providers/azure-speech.md for the\n  Azure-MP3 case.\nSTT_AUDIO_TOO_LONG is thrown when the buffer exceeds the per-call\n  maxAudioBytes limit. Default is 25 MB (matches Whisper's documented\n  ceiling). Override via stt.maxAudioBytes.","hierarchy":{"lvl0":"Reference","lvl1":"Error Code Reference","lvl2":"Common Triggers","lvl3":""}},
 {"objectID":"df1592c1b3ccdbed96fbbc998d2facc217e2652c2a3cbb2d630ed64f1c85fdb9","title":"Related Documentation","url":"/docs/reference/error-codes#related-documentation","content":"Troubleshooting Guide - Common issues and solutions\nConfiguration Reference - Environment variables and settings\nFAQ - Frequently asked questions\nProvider Feature Compatibility - Provider capabilities matrix","hierarchy":{"lvl0":"Reference","lvl1":"Error Code Reference","lvl2":"Related Documentation","lvl3":""}},
-{"objectID":"4dccb885a61a05030e1eeb0b2413af7062a282bcce434d742561026c5af1d015","title":"Frequently Asked Questions","url":"/docs/reference/faq","content":"Frequently Asked Questions\n\nCommon questions and answers about NeuroLink usage, configuration, and troubleshooting.\n\n🚀 Getting Started\n\nQ: What is NeuroLink?\n\nA: NeuroLink is an enterprise AI development platform that provides unified access to multiple AI providers (OpenAI, Google AI, Anthropic, AWS Bedrock, etc.) through a single SDK and CLI. It includes built-in tools, analytics, evaluation capabilities, and supports the Model Context Protocol (MCP) for extended functionality.\n\nQ: Which AI providers does NeuroLink support?\n\nA: NeuroLink ships AI providers for text generation, streaming, and decision-making — plus separate provider systems for voice and media generation. The text and multimodal providers include:\nOpenAI (GPT-4o, GPT-4.1, o3, o4-mini)\nGoogle AI Studio (Gemini 3 Flash/Pro, Gemini 2.5 Pro/Flash)\nGoogle Vertex AI (Gemini 3, Claude via Vertex)\nAnthropic (Claude Opus 4.7, Sonnet 4.6, 4.5 Opus/Sonnet/Haiku)\nAWS Bedrock (Claude, Titan, Nova models)\nAzure OpenAI (GPT models)\nHugging Face (Open source models)\nOllama (Local AI models)\nMistral AI (Mistral models)\nLiteLLM (100+ models via proxy)\nAWS SageMaker (Custom endpoints)\nOpenAI-compatible (Any OpenAI-API-compatible endpoint)\nOpenRouter (300+ models via OpenRouter)\nDeepSeek (DeepSeek V3, R1)\nNVIDIA NIM (Llama 3.3 70B, 400+ catalog models)\nLM Studio (Local models loaded in LM Studio)\nllama.cpp (Local GGUF models via llama-server)\nGroq, Cerebras, SambaNova, Together AI, Fireworks AI, Perplexity, Cloudflare Workers AI, xAI, Baseten, GMI Cloud, Inception Labs, io.net Intelligence, Mancer, Upstage, API Route (zero-quirk OpenAI-wire-compatible catalog providers)\nCohere (chat, plus embed() and reranking)\nVoyage AI, Jina AI (embedding and/or reranking only — no chat completions)\nTypeSafe Jev, Laya, XOR, SubQ, Perplexity Decisions, Cloudflare Clef, OpenAI Decisions (decision-only — serve decide(), not generate()/stream(); tried in the order TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef, OpenAI Decisions when several are configured; Clef is configured by the same CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID as the Cloudflare Workers AI text provider, and OpenAI Decisions by the same OPENAI_API_KEY as the OpenAI chat provider, except that this key is not used by default while OPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com, since it is then a gateway's)\n\nSee Provider Setup for the complete roster with setup guides.\n\nVoice providers (a separate system from the providers above):\nOpenAI TTS (TTS-1, TTS-1-HD, GPT-4o Audio)\nElevenLabs (Multilingual v2, Turbo v2.5, Flash v2.5)\nDeepgram (Nova-3, Nova-2, Enhanced — STT)\nAzure Speech (Azure Cognitive Services TTS + STT)\nGoogle TTS / STT (Google Cloud Speech)\nWhisper (OpenAI Whisper — STT)\nFish Audio (TTS)\nCartesia (TTS)\nOpenAI Realtime + Gemini Live (realtime voice APIs)\n\nMedia generation providers (image / video / music / avatar) — Kling, Runway, Replicate, Beatoven, Lyria, D-ID, HeyGen. See Media Generation for the full list.\n\nQ: Do I need to install anything?\n\nA: No installation required! You can use NeuroLink directly with npx:\n\nFor frequent use, you can install globally: npm install -g @juspay/neurolink\n\n🔧 Configuration\n\nQ: How do I set up API keys?\n\nA: Create a .env file in your project directory:\n\nNeuroLink automatically loads these environment variables.\n\nQ: Can I use NeuroLink behind a corporate proxy?\n\nA: Yes! NeuroLink automatically detects and uses corporate proxy settings:\n\nNo additional configuration needed.\n\nQ: How do I configure multiple environments (dev/staging/prod)?\n\nA: Use environment-specific .env files:\n\n🎯 Usage\n\nQ: What's the difference between CLI and SDK?\n\nA:\n\n| Feature              | CLI                          | SDK                       |\n| -------------------- | ---------------------------- | ------------------------- |\n| Best for         | Scripts, automation, testing | Applications, integration |\n| Installation     | None required (npx)          | npm install required      |\n| Output           | Text, JSON                   | Native JavaScript objects |\n| Batch processing | Built-in batch command     | Manual implementation     |\n| Learning curve   | Low                          | Medium                    |\n\nQ: How do I choose the best provider for my use case?\n\nA: NeuroLink can auto-select the best provider, or you can choose based on:\nSpeed: Google AI (fastest responses)\nCoding: Anthropic Claude (best for code analysis)\nCreative: OpenAI (best for creative content)\nCost: Google AI Studio (free tier available)\nEnterprise: AWS Bedrock or Azure OpenAI\n\nQ: Can I use multiple providers in the same application?\n\nA: Yes! You can specify different providers for different requests:\n\n🔍 Troubleshooting\n\nQ: Why am I getting \"API key not found\" errors?\n\nA: Common solutions:\nCheck .env file exists and is in the correct directory\nVerify file format: No spaces around = signs\nCheck file permissions: .env file should be read","hierarchy":{"lvl0":"Reference","lvl1":"Frequently Asked Questions","lvl2":"","lvl3":""}},
+{"objectID":"4dccb885a61a05030e1eeb0b2413af7062a282bcce434d742561026c5af1d015","title":"Frequently Asked Questions","url":"/docs/reference/faq","content":"Frequently Asked Questions\n\nCommon questions and answers about NeuroLink usage, configuration, and troubleshooting.\n\n🚀 Getting Started\n\nQ: What is NeuroLink?\n\nA: NeuroLink is an enterprise AI development platform that provides unified access to multiple AI providers (OpenAI, Google AI, Anthropic, AWS Bedrock, etc.) through a single SDK and CLI. It includes built-in tools, analytics, evaluation capabilities, and supports the Model Context Protocol (MCP) for extended functionality.\n\nQ: Which AI providers does NeuroLink support?\n\nA: NeuroLink ships AI providers for text generation, streaming, and decision-making — plus separate provider systems for voice and media generation. The text and multimodal providers include:\nOpenAI (GPT-4o, GPT-4.1, o3, o4-mini)\nGoogle AI Studio (Gemini 3 Flash/Pro, Gemini 2.5 Pro/Flash)\nGoogle Vertex AI (Gemini 3, Claude via Vertex)\nAnthropic (Claude Opus 4.7, Sonnet 4.6, 4.5 Opus/Sonnet/Haiku)\nAWS Bedrock (Claude, Titan, Nova models)\nAzure OpenAI (GPT models)\nHugging Face (Open source models)\nOllama (Local AI models)\nMistral AI (Mistral models)\nLiteLLM (100+ models via proxy)\nAWS SageMaker (Custom endpoints)\nOpenAI-compatible (Any OpenAI-API-compatible endpoint)\nOpenRouter (300+ models via OpenRouter)\nDeepSeek (DeepSeek V3, R1)\nNVIDIA NIM (Llama 3.3 70B, 400+ catalog models)\nLM Studio (Local models loaded in LM Studio)\nllama.cpp (Local GGUF models via llama-server)\nGroq, Cerebras, SambaNova, Together AI, Fireworks AI, Perplexity, Cloudflare Workers AI, xAI, Baseten, GMI Cloud, Inception Labs, io.net Intelligence, Mancer, Upstage, API Route (zero-quirk OpenAI-wire-compatible catalog providers)\nCohere (chat, plus embed() and reranking)\nVoyage AI, Jina AI (embedding and/or reranking only — no chat completions)\nTypeSafe Jev, Laya, XOR, SubQ, Perplexity Decisions, Cloudflare Clef, OpenAI Decisions, ScaleDown (decision-only — serve decide(), not generate()/stream(); tried in the order TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef, OpenAI Decisions, ScaleDown when several are configured; Clef is configured by the same CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID as the Cloudflare Workers AI text provider, and OpenAI Decisions by the same OPENAI_API_KEY as the OpenAI chat provider, except that this key is not used by default while OPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com, since it is then a gateway's; ScaleDown is configured by its own SCALEDOWN_API_KEY, which no other provider reads)\n\nSee Provider Setup for the complete roster with setup guides.\n\nVoice providers (a separate system from the providers above):\nOpenAI TTS (TTS-1, TTS-1-HD, GPT-4o Audio)\nElevenLabs (Multilingual v2, Turbo v2.5, Flash v2.5)\nDeepgram (Nova-3, Nova-2, Enhanced — STT)\nAzure Speech (Azure Cognitive Services TTS + STT)\nGoogle TTS / STT (Google Cloud Speech)\nWhisper (OpenAI Whisper — STT)\nFish Audio (TTS)\nCartesia (TTS)\nOpenAI Realtime + Gemini Live (realtime voice APIs)\n\nMedia generation providers (image / video / music / avatar) — Kling, Runway, Replicate, Beatoven, Lyria, D-ID, HeyGen. See Media Generation for the full list.\n\nQ: Do I need to install anything?\n\nA: No installation required! You can use NeuroLink directly with npx:\n\nFor frequent use, you can install globally: npm install -g @juspay/neurolink\n\n🔧 Configuration\n\nQ: How do I set up API keys?\n\nA: Create a .env file in your project directory:\n\nNeuroLink automatically loads these environment variables.\n\nQ: Can I use NeuroLink behind a corporate proxy?\n\nA: Yes! NeuroLink automatically detects and uses corporate proxy settings:\n\nNo additional configuration needed.\n\nQ: How do I configure multiple environments (dev/staging/prod)?\n\nA: Use environment-specific .env files:\n\n🎯 Usage\n\nQ: What's the difference between CLI and SDK?\n\nA:\n\n| Feature              | CLI                          | SDK                       |\n| -------------------- | ---------------------------- | ------------------------- |\n| Best for         | Scripts, automation, testing | Applications, integration |\n| Installation     | None required (npx)          | npm install required      |\n| Output           | Text, JSON                   | Native JavaScript objects |\n| Batch processing | Built-in batch command     | Manual implementation     |\n| Learning curve   | Low                          | Medium                    |\n\nQ: How do I choose the best provider for my use case?\n\nA: NeuroLink can auto-select the best provider, or you can choose based on:\nSpeed: Google AI (fastest responses)\nCoding: Anthropic Claude (best for code analysis)\nCreative: OpenAI (best for creative content)\nCost: Google AI Studio (free tier available)\nEnterprise: AWS Bedrock or Azure OpenAI\n\nQ: Can I use multiple providers in the same application?\n\nA: Yes! You can specify different providers for different requests:\n\n🔍 Troubleshooting\n\nQ: Why am I getting \"API key not found\" errors?\n\nA: Common solutions:\nCheck .env file exists and is in the corr","hierarchy":{"lvl0":"Reference","lvl1":"Frequently Asked Questions","lvl2":"","lvl3":""}},
 {"objectID":"430e71707f7db7f9866d7ab0b23f14f769c4812334161d853920bf4358ad8c17","title":"Frequently Asked Questions","url":"/docs/reference/faq#frequently-asked-questions","content":"Common questions and answers about NeuroLink usage, configuration, and troubleshooting.","hierarchy":{"lvl0":"Reference","lvl1":"Frequently Asked Questions","lvl2":"Frequently Asked Questions","lvl3":""}},
 {"objectID":"e9218a0cd45a1960da2da4d8d664f5a8b77494352527c7b4d1e6bc12c137400f","title":"Q: What is NeuroLink?","url":"/docs/reference/faq#q-what-is-neurolink","content":"A: NeuroLink is an enterprise AI development platform that provides unified access to multiple AI providers (OpenAI, Google AI, Anthropic, AWS Bedrock, etc.) through a single SDK and CLI. It includes built-in tools, analytics, evaluation capabilities, and supports the Model Context Protocol (MCP) for extended functionality.","hierarchy":{"lvl0":"Reference","lvl1":"Frequently Asked Questions","lvl2":"Q: What is NeuroLink?","lvl3":""}},
-{"objectID":"c32f5558167b5817bde6577d51c68bbce793c3118b1192fdbf6c6e5958d3faff","title":"Q: Which AI providers does NeuroLink support?","url":"/docs/reference/faq#q-which-ai-providers-does-neurolink-support","content":"A: NeuroLink ships AI providers for text generation, streaming, and decision-making — plus separate provider systems for voice and media generation. The text and multimodal providers include:\nOpenAI (GPT-4o, GPT-4.1, o3, o4-mini)\nGoogle AI Studio (Gemini 3 Flash/Pro, Gemini 2.5 Pro/Flash)\nGoogle Vertex AI (Gemini 3, Claude via Vertex)\nAnthropic (Claude Opus 4.7, Sonnet 4.6, 4.5 Opus/Sonnet/Haiku)\nAWS Bedrock (Claude, Titan, Nova models)\nAzure OpenAI (GPT models)\nHugging Face (Open source models)\nOllama (Local AI models)\nMistral AI (Mistral models)\nLiteLLM (100+ models via proxy)\nAWS SageMaker (Custom endpoints)\nOpenAI-compatible (Any OpenAI-API-compatible endpoint)\nOpenRouter (300+ models via OpenRouter)\nDeepSeek (DeepSeek V3, R1)\nNVIDIA NIM (Llama 3.3 70B, 400+ catalog models)\nLM Studio (Local models loaded in LM Studio)\nllama.cpp (Local GGUF models via llama-server)\nGroq, Cerebras, SambaNova, Together AI, Fireworks AI, Perplexity, Cloudflare Workers AI, xAI, Baseten, GMI Cloud, Inception Labs, io.net Intelligence, Mancer, Upstage, API Route (zero-quirk OpenAI-wire-compatible catalog providers)\nCohere (chat, plus embed() and reranking)\nVoyage AI, Jina AI (embedding and/or reranking only — no chat completions)\nTypeSafe Jev, Laya, XOR, SubQ, Perplexity Decisions, Cloudflare Clef, OpenAI Decisions (decision-only — serve decide(), not generate()/stream(); tried in the order TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef, OpenAI Decisions when several are configured; Clef is configured by the same CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID as the Cloudflare Workers AI text provider, and OpenAI Decisions by the same OPENAI_API_KEY as the OpenAI chat provider, except that this key is not used by default while OPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com, since it is then a gateway's)\n\nSee Provider Setup for the complete roster with setup guides.","hierarchy":{"lvl0":"Reference","lvl1":"Frequently Asked Questions","lvl2":"Q: Which AI providers does NeuroLink support?","lvl3":""}},
+{"objectID":"c32f5558167b5817bde6577d51c68bbce793c3118b1192fdbf6c6e5958d3faff","title":"Q: Which AI providers does NeuroLink support?","url":"/docs/reference/faq#q-which-ai-providers-does-neurolink-support","content":"A: NeuroLink ships AI providers for text generation, streaming, and decision-making — plus separate provider systems for voice and media generation. The text and multimodal providers include:\nOpenAI (GPT-4o, GPT-4.1, o3, o4-mini)\nGoogle AI Studio (Gemini 3 Flash/Pro, Gemini 2.5 Pro/Flash)\nGoogle Vertex AI (Gemini 3, Claude via Vertex)\nAnthropic (Claude Opus 4.7, Sonnet 4.6, 4.5 Opus/Sonnet/Haiku)\nAWS Bedrock (Claude, Titan, Nova models)\nAzure OpenAI (GPT models)\nHugging Face (Open source models)\nOllama (Local AI models)\nMistral AI (Mistral models)\nLiteLLM (100+ models via proxy)\nAWS SageMaker (Custom endpoints)\nOpenAI-compatible (Any OpenAI-API-compatible endpoint)\nOpenRouter (300+ models via OpenRouter)\nDeepSeek (DeepSeek V3, R1)\nNVIDIA NIM (Llama 3.3 70B, 400+ catalog models)\nLM Studio (Local models loaded in LM Studio)\nllama.cpp (Local GGUF models via llama-server)\nGroq, Cerebras, SambaNova, Together AI, Fireworks AI, Perplexity, Cloudflare Workers AI, xAI, Baseten, GMI Cloud, Inception Labs, io.net Intelligence, Mancer, Upstage, API Route (zero-quirk OpenAI-wire-compatible catalog providers)\nCohere (chat, plus embed() and reranking)\nVoyage AI, Jina AI (embedding and/or reranking only — no chat completions)\nTypeSafe Jev, Laya, XOR, SubQ, Perplexity Decisions, Cloudflare Clef, OpenAI Decisions, ScaleDown (decision-only — serve decide(), not generate()/stream(); tried in the order TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef, OpenAI Decisions, ScaleDown when several are configured; Clef is configured by the same CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID as the Cloudflare Workers AI text provider, and OpenAI Decisions by the same OPENAI_API_KEY as the OpenAI chat provider, except that this key is not used by default while OPENAI_BASE_URL, OPENAI_API_BASE or credentials.openai.baseURL names a host other than https://api.openai.com, since it is then a gateway's; ScaleDown is configured by its own SCALEDOWN_API_KEY, which no other provider reads)\n\nSee","hierarchy":{"lvl0":"Reference","lvl1":"Frequently Asked Questions","lvl2":"Q: Which AI providers does NeuroLink support?","lvl3":""}},
 {"objectID":"c7da7431322332ad0e9e6439acb89a8a9f6af428d15d84c94adeb886ea4fdaab","title":"Q: Do I need to install anything?","url":"/docs/reference/faq#q-do-i-need-to-install-anything","content":"A: No installation required! You can use NeuroLink directly with npx:\n\nFor frequent use, you can install globally: npm install -g @juspay/neurolink","hierarchy":{"lvl0":"Reference","lvl1":"Frequently Asked Questions","lvl2":"Q: Do I need to install anything?","lvl3":""}},
 {"objectID":"379b5c908063649275668a696f71c0e87923dd0d011f221a19a43f69a483acea","title":"Q: How do I set up API keys?","url":"/docs/reference/faq#q-how-do-i-set-up-api-keys","content":"A: Create a .env file in your project directory:\n\n`bash","hierarchy":{"lvl0":"Reference","lvl1":"Frequently Asked Questions","lvl2":"Q: How do I set up API keys?","lvl3":""}},
 {"objectID":"07c128fc6b5bcea9361ff3423a97e6a5a4e85fb14571f4a5d14bcb11eb389105","title":".env file","url":"/docs/reference/faq#env-file","content":"OPENAIAPIKEY=\"sk-your-openai-key\"\nGOOGLEAIAPI_KEY=\"AIza-your-google-ai-key\"\nANTHROPICAPIKEY=\"sk-ant-your-anthropic-key\"","hierarchy":{"lvl0":"Reference","lvl1":"Frequently Asked Questions","lvl2":".env file","lvl3":""}},
@@ -10530,7 +10558,7 @@
 {"objectID":"ca34f307c204381b1904fe62e5762d2451f439c5f2c8d47083804d9e9ca1f361","title":"Factory Functions","url":"/docs/reference/server-configuration#factory-functions","content":"For quick OpenAPI generation without instantiating the class:","hierarchy":{"lvl0":"Reference","lvl1":"Srvr Cofiguratio Rfrc []","lvl2":"Factory Functions","lvl3":""}},
 {"objectID":"8e5b75ed2dd85636a125c03319a725a2c10669870a925448135351aa81e49df9","title":"All Available Schemas","url":"/docs/reference/server-configuration#all-available-schemas","content":"The OpenAPISchemas registry provides access to all built-in schemas:","hierarchy":{"lvl0":"Reference","lvl1":"Srvr Cofiguratio Rfrc []","lvl2":"All Available Schemas","lvl3":""}},
 {"objectID":"50b633e0ff979961fac99435d14bff5b692d058757f69642edd9efc0db8c050f","title":"Related Documentation","url":"/docs/reference/server-configuration#related-documentation","content":"Server Adapters Overview - Introduction to server adapters\nSecurity Guide - Security best practices\nDeployment Guide - Deployment strategies and configurations","hierarchy":{"lvl0":"Reference","lvl1":"Srvr Cofiguratio Rfrc []","lvl2":"Related Documentation","lvl3":""}},
-{"objectID":"efd7e891fd4b311148a6006fb8a03a36f7702832577266f8e750ec6d79d4dd25","title":"NeuroLink Troubleshooting Guide","url":"/docs/reference/troubleshooting","content":"NeuroLink Troubleshooting Guide\n\nVersion: v9.26.1\nLast Updated: March 2026\n\nOverview\n\nThis guide helps diagnose and resolve common issues with NeuroLink, including AI provider connectivity, MCP integration, CLI usage problems, streaming issues, and the generate function migration.\n\nQuick Diagnostics\n\nBefore diving into specific issues, try these quick diagnostics:\n\nQuick Fixes\n\n| Symptom                                | Resolution                                                                                                                     |\n| -------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------ |\n| Image not found when using --image | Provide an absolute path or run the command from the directory containing the asset. URLs must be HTTPS.                       |\n| Evaluation model not configured      | Set NEUROLINK_EVALUATION_PROVIDER/NEUROLINK_EVALUATION_MODEL, or disable --enableEvaluation until credentials are added. |\n| Redis connection failed in loop mode | Export REDIS_URL before running neurolink loop or start the session with --no-auto-redis.                                |\n| Model not available in region        | Confirm the model supports the requested region and update AWS_REGION / GOOGLE_VERTEX_LOCATION accordingly.                |\n| CLI exits after error inside loop      | Upgrade to latest @juspay/neurolink and restart the loop; new builds catch errors without exiting.                           |\n\nQ4 2025 Features -- Common Issues\n\nHuman-in-the-Loop (HITL)\n\n| Issue                                   | Solution                                                                                                    |\n| --------------------------------------- | ----------------------------------------------------------------------------------------------------------- |\n| Tool executes without asking permission | Add requiresConfirmation: true to tool definition. See HITL Guide    |\n| Confirmation dialog doesn't appear      | Handle USER_CONFIRMATION_REQUIRED error in your UI. See HITL Guide |\n| Permission flag not resetting           | Call setUserConfirmation(false) after tool execution. See HITL Guide  |\n\nGuardrails Middleware\n\n| Issue                      | Solution                                                                                                               |\n| -------------------------- | ---------------------------------------------------------------------------------------------------------------------- |\n| Content not being filtered | Ensure preset: \"security\" is set in middleware config. See Guardrails Guide |\n| Too many false positives   | Review bad word list, remove common words. See Guardrails Guide            |\n| Model-based filter is slow | Switch to gpt-4o-mini for faster filtering. See Guardrails Guide        |\n\nRedis Conversation Export\n\n| Issue                                        | Solution                                                                                                                                     |\n| -------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------- |\n| Export returns empty history                 | Verify Redis connection and session ID exists. See Conversation History Guide         |\n| getConversationHistory returns empty array | Ensure conversationMemory.enabled: true is configured. See Conversation History Guide |\n| Missing metadata in export                   | Set includeMetadata: true in export options. See Conversation History Guide          |\n\nVideo Generation (Veo 3.1)\n\n| Issue                                  | Solution                                                                                                                                                        |\n| -------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| PROVIDER_NOT_CONFIGURED error        | Set GOOGLE_APPLICATION_CREDENTIALS to your service account JSON path. See Video Generation Guide             |\n| VIDEO_POLL_TIMEOUT after 3 minutes   | Video generation can take 1-2 minutes; increase timeout or check Vertex AI quota. See Video Generation Guide |\n| VIDEO_INVALID_INPUT for image format | Ensure image is PNG, JPEG, or WebP under 20MB; check aspect ratio compatibility. See Video Generation Guide      |\n| Video generation uses wrong provider   | Video gen only supports Vertex AI; provider auto-switches to vertex when output.mode: \"video\"                                                               |\n| Project not found error              | Set GOOGLE_VERTEX_PROJECT or GOOGLE_CLOUD_PROJECT environment variable                                                ","hierarchy":{"lvl0":"Reference","lvl1":"NeuroLink Troubleshooting Guide","lvl2":"","lvl3":""}},
+{"objectID":"efd7e891fd4b311148a6006fb8a03a36f7702832577266f8e750ec6d79d4dd25","title":"NeuroLink Troubleshooting Guide","url":"/docs/reference/troubleshooting","content":"NeuroLink Troubleshooting Guide\n\nVersion: v9.26.1\nLast Updated: March 2026\n\nOverview\n\nThis guide helps diagnose and resolve common issues with NeuroLink, including AI provider connectivity, MCP integration, CLI usage problems, streaming issues, and the generate function migration.\n\nQuick Diagnostics\n\nBefore diving into specific issues, try these quick diagnostics:\n\nQuick Fixes\n\n| Symptom                                | Resolution                                                                                                                     |\n| -------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------ |\n| Image not found when using --image | Provide an absolute path or run the command from the directory containing the asset. URLs must be HTTPS.                       |\n| Evaluation model not configured      | Set NEUROLINK_EVALUATION_PROVIDER/NEUROLINK_EVALUATION_MODEL, or disable --enableEvaluation until credentials are added. |\n| Redis connection failed in loop mode | Export REDIS_URL before running neurolink loop or start the session with --no-auto-redis.                                |\n| Model not available in region        | Confirm the model supports the requested region and update AWS_REGION / GOOGLE_VERTEX_LOCATION accordingly.                |\n| CLI exits after error inside loop      | Upgrade to latest @juspay/neurolink and restart the loop; new builds catch errors without exiting.                           |\n\nQ4 2025 Features -- Common Issues\n\nHuman-in-the-Loop (HITL)\n\n| Issue                                   | Solution                                                                                                    |\n| --------------------------------------- | ----------------------------------------------------------------------------------------------------------- |\n| Tool executes without asking permission | Add requiresConfirmation: true to tool definition. See HITL Guide    |\n| Confirmation dialog doesn't appear      | Handle USER_CONFIRMATION_REQUIRED error in your UI. See HITL Guide |\n| Permission flag not resetting           | Call setUserConfirmation(false) after tool execution. See HITL Guide  |\n\nGuardrails Middleware\n\n| Issue                      | Solution                                                                                                                                                                                                            |\n| -------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Content not being filtered | preset: \"security\" only turns guardrails on; also set badWords (with enabled: true and a list) or modelFilter. See Guardrails Guide |\n| Too many false positives   | Review bad word list, remove common words. See Guardrails Guide                                                                                                         |\n| Model-based filter is slow | Switch to openai:gpt-4o-mini for faster filtering. See Guardrails Guide                                                                                              |\n\nRedis Conversation Export\n\n| Issue                                        | Solution                                                                                                                                     |\n| -------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------- |\n| Export returns empty history                 | Verify Redis connection and session ID exists. See Conversation History Guide         |\n| getConversationHistory returns empty array | Ensure conversationMemory.enabled: true is configured. See Conversation History Guide |\n| Missing metadata in export                   | Set includeMetadata: true in export options. See Conversation History Guide          |\n\nVideo Generation (Veo 3.1)\n\n| Issue                                  | Solution                                                                                                                                                        |\n| -------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| PROVIDER_NOT_CONFIGURED error        | Set GOOGLE_APPLICATION_CREDENTIALS to your service account JSON path. See Video Generation Guide             |\n| VIDEO_POLL_TIMEOUT after 3 minutes   | Video generation can take 1-2 minutes; increase timeout or check Vertex AI quota. See Video Generation Guide |\n| VIDEO_INVALID_INPUT for image format | Ensure image is PNG, JPEG, or WebP under 20MB; ","hierarchy":{"lvl0":"Reference","lvl1":"NeuroLink Troubleshooting Guide","lvl2":"","lvl3":""}},
 {"objectID":"f18cd128c5fcffdd4422033accb929756a0ed04e990d54c0915a2270fc99b7bb","title":"NeuroLink Troubleshooting Guide","url":"/docs/reference/troubleshooting#neurolink-troubleshooting-guide","content":"Version: v9.26.1\nLast Updated: March 2026","hierarchy":{"lvl0":"Reference","lvl1":"NeuroLink Troubleshooting Guide","lvl2":"NeuroLink Troubleshooting Guide","lvl3":""}},
 {"objectID":"16da6024adabd6a43133f9085cea031c955dbe961369e449b1e36739585501f9","title":"Overview","url":"/docs/reference/troubleshooting#overview","content":"This guide helps diagnose and resolve common issues with NeuroLink, including AI provider connectivity, MCP integration, CLI usage problems, streaming issues, and the generate function migration.","hierarchy":{"lvl0":"Reference","lvl1":"NeuroLink Troubleshooting Guide","lvl2":"Overview","lvl3":""}},
 {"objectID":"989cb831fb47ad94f3f9469ca199da828c0d637b2acc3d200ffea2a2a3ae563a","title":"Quick Diagnostics","url":"/docs/reference/troubleshooting#quick-diagnostics","content":"Before diving into specific issues, try these quick diagnostics:\n\n`bash","hierarchy":{"lvl0":"Reference","lvl1":"NeuroLink Troubleshooting Guide","lvl2":"Quick Diagnostics","lvl3":""}},
@@ -10542,7 +10570,7 @@
 {"objectID":"21f39a157951a67d608fc7cdf36c85389d8ff578c25e8d2293ab7cf5a9942e3b","title":"6. Enable debug logging","url":"/docs/reference/troubleshooting#6-enable-debug-logging","content":"npx @juspay/neurolink generate \"Test\" --debug\n`","hierarchy":{"lvl0":"Reference","lvl1":"NeuroLink Troubleshooting Guide","lvl2":"6. Enable debug logging","lvl3":""}},
 {"objectID":"9875cf71575de93b4835f39b84e856f4a94df8899571e70f3dfc8336d16f69d7","title":"Quick Fixes","url":"/docs/reference/troubleshooting#quick-fixes","content":"| Symptom                                | Resolution                                                                                                                     |\n| -------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------ |\n| Image not found when using --image | Provide an absolute path or run the command from the directory containing the asset. URLs must be HTTPS.                       |\n| Evaluation model not configured      | Set NEUROLINK_EVALUATION_PROVIDER/NEUROLINK_EVALUATION_MODEL, or disable --enableEvaluation until credentials are added. |\n| Redis connection failed in loop mode | Export REDIS_URL before running neurolink loop or start the session with --no-auto-redis.                                |\n| Model not available in region        | Confirm the model supports the requested region and update AWS_REGION / GOOGLE_VERTEX_LOCATION accordingly.                |\n| CLI exits after error inside loop      | Upgrade to latest @juspay/neurolink and restart the loop; new builds catch errors without exiting.                           |","hierarchy":{"lvl0":"Reference","lvl1":"NeuroLink Troubleshooting Guide","lvl2":"Quick Fixes","lvl3":""}},
 {"objectID":"519835175f0f4ba411dbe92426eeb603aa23a38ec149913ac19bca7682b3adec","title":"Human-in-the-Loop (HITL)","url":"/docs/reference/troubleshooting#human-in-the-loop-hitl","content":"| Issue                                   | Solution                                                                                                    |\n| --------------------------------------- | ----------------------------------------------------------------------------------------------------------- |\n| Tool executes without asking permission | Add requiresConfirmation: true to tool definition. See HITL Guide    |\n| Confirmation dialog doesn't appear      | Handle USER_CONFIRMATION_REQUIRED error in your UI. See HITL Guide |\n| Permission flag not resetting           | Call setUserConfirmation(false) after tool execution. See HITL Guide  |","hierarchy":{"lvl0":"Reference","lvl1":"NeuroLink Troubleshooting Guide","lvl2":"Human-in-the-Loop (HITL)","lvl3":""}},
-{"objectID":"3e08d9134df8755f7ff0cd9382ee0a4151ff1a1ff6c13f60494e8b5206dea416","title":"Guardrails Middleware","url":"/docs/reference/troubleshooting#guardrails-middleware","content":"| Issue                      | Solution                                                                                                               |\n| -------------------------- | ---------------------------------------------------------------------------------------------------------------------- |\n| Content not being filtered | Ensure preset: \"security\" is set in middleware config. See Guardrails Guide |\n| Too many false positives   | Review bad word list, remove common words. See Guardrails Guide            |\n| Model-based filter is slow | Switch to gpt-4o-mini for faster filtering. See Guardrails Guide        |","hierarchy":{"lvl0":"Reference","lvl1":"NeuroLink Troubleshooting Guide","lvl2":"Guardrails Middleware","lvl3":""}},
+{"objectID":"3e08d9134df8755f7ff0cd9382ee0a4151ff1a1ff6c13f60494e8b5206dea416","title":"Guardrails Middleware","url":"/docs/reference/troubleshooting#guardrails-middleware","content":"| Issue                      | Solution                                                                                                                                                                                                            |\n| -------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Content not being filtered | preset: \"security\" only turns guardrails on; also set badWords (with enabled: true and a list) or modelFilter. See Guardrails Guide |\n| Too many false positives   | Review bad word list, remove common words. See Guardrails Guide                                                                                                         |\n| Model-based filter is slow | Switch to openai:gpt-4o-mini for faster filtering. See Guardrails Guide                                                                                              |","hierarchy":{"lvl0":"Reference","lvl1":"NeuroLink Troubleshooting Guide","lvl2":"Guardrails Middleware","lvl3":""}},
 {"objectID":"13407b41e672f252dd093443b23887d95a5d464a743ab992713dc63bbcf42cc2","title":"Redis Conversation Export","url":"/docs/reference/troubleshooting#redis-conversation-export","content":"| Issue                                        | Solution                                                                                                                                     |\n| -------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------- |\n| Export returns empty history                 | Verify Redis connection and session ID exists. See Conversation History Guide         |\n| getConversationHistory returns empty array | Ensure conversationMemory.enabled: true is configured. See Conversation History Guide |\n| Missing metadata in export                   | Set includeMetadata: true in export options. See Conversation History Guide          |","hierarchy":{"lvl0":"Reference","lvl1":"NeuroLink Troubleshooting Guide","lvl2":"Redis Conversation Export","lvl3":""}},
 {"objectID":"6cf2bc5f5079c8e4490e3d9a2b8d687e06ef94abedb94924e0e92788ed85256a","title":"Video Generation (Veo 3.1)","url":"/docs/reference/troubleshooting#video-generation-veo-31","content":"| Issue                                  | Solution                                                                                                                                                        |\n| -------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| PROVIDER_NOT_CONFIGURED error        | Set GOOGLE_APPLICATION_CREDENTIALS to your service account JSON path. See Video Generation Guide             |\n| VIDEO_POLL_TIMEOUT after 3 minutes   | Video generation can take 1-2 minutes; increase timeout or check Vertex AI quota. See Video Generation Guide |\n| VIDEO_INVALID_INPUT for image format | Ensure image is PNG, JPEG, or WebP under 20MB; check aspect ratio compatibility. See Video Generation Guide      |\n| Video generation uses wrong provider   | Video gen only supports Vertex AI; provider auto-switches to vertex when output.mode: \"video\"                                                               |\n| Project not found error              | Set GOOGLE_VERTEX_PROJECT or GOOGLE_CLOUD_PROJECT environment variable                                                                                      |\n| Audio missing from generated video     | Set output.video.audio: true (enabled by default) and ensure Veo 3.1 model is used                                                                            |","hierarchy":{"lvl0":"Reference","lvl1":"NeuroLink Troubleshooting Guide","lvl2":"Video Generation (Veo 3.1)","lvl3":""}},
 {"objectID":"a92a4e02183ff63c57c4e6477e15e48f9bc4b2ae17fd61849e537bd592dad0d9","title":"PPT Generation (PowerPoint Presentations)","url":"/docs/reference/troubleshooting#ppt-generation-powerpoint-presentations","content":"PPT_PLANNING_FAILED -- Check AI provider connection and ensure valid prompt. See PPT Generation Guide\nPPT_INVALID_AI_RESPONSE during generation -- Simplify prompt/topic and retry. See PPT Generation Guide\nPPT_FILE_WRITE_FAILED -- Check write permissions for output directory and disk space. See PPT Generation Guide\nEmpty slides in presentation -- Ensure content plan has enough detail; try more specific prompts\nImages not generating -- Set generateAIImages: true in output.ppt (SDK) or avoid --pptNoImages (CLI), and configure VERTEX_IMAGE_MODEL. See PPT Generation Guide\nTheme not applying correctly -- Verify theme name: modern, corporate, creative, minimal, or dark","hierarchy":{"lvl0":"Reference","lvl1":"NeuroLink Troubleshooting Guide","lvl2":"PPT Generation (PowerPoint Presentations)","lvl3":""}},
@@ -11008,7 +11036,7 @@
 {"objectID":"357bbb718f458d764e00524826a1d42a64da83167df93a52c04d9ddec3651764","title":"Core Events","url":"/docs/sdk/api-reference#core-events","content":"| Event              | Emitted When                         |\n| ------------------ | ------------------------------------ |\n| generation:start | A generate() call begins           |\n| generation:end   | A generate() call completes        |\n| stream:start     | A stream() call begins             |\n| stream:chunk     | A new chunk arrives during streaming |\n| stream:end       | Streaming completes normally         |\n| stream:complete  | Stream fully consumed                |\n| stream:error     | An error occurs during streaming     |\n| tool:start       | A tool execution begins              |\n| tool:end         | A tool execution completes           |\n| response:start   | A provider response begins           |\n| response:end     | A provider response completes        |","hierarchy":{"lvl0":"Sdk","lvl1":"API Reference","lvl2":"Core Events","lvl3":""}},
 {"objectID":"efcf19aa651c13dbc72ef4b3d0f83986ee180008a076c0db030509880e4d4b1a","title":"MCP Server Events","url":"/docs/sdk/api-reference#mcp-server-events","content":"| Event                            | Emitted When                              |\n| -------------------------------- | ----------------------------------------- |\n| externalMCP:serverConnected    | An external MCP server connects           |\n| externalMCP:serverDisconnected | An external MCP server disconnects        |\n| externalMCP:serverFailed       | An external MCP server connection fails   |\n| externalMCP:toolDiscovered     | A new tool is discovered on an MCP server |\n| externalMCP:toolRemoved        | A tool is removed from an MCP server      |\n| externalMCP:serverAdded        | A new MCP server registration is added    |\n| externalMCP:serverRemoved      | An MCP server registration is removed     |","hierarchy":{"lvl0":"Sdk","lvl1":"API Reference","lvl2":"MCP Server Events","lvl3":""}},
 {"objectID":"c4d157804d0d96f1d8571e6aabeff4305aca0c07c151894dee3804d611331de6","title":"Other Events","url":"/docs/sdk/api-reference#other-events","content":"| Event                  | Emitted When                        |\n| ---------------------- | ----------------------------------- |\n| tools-register:start | Tool registration process begins    |\n| tools-register:end   | Tool registration process completes |\n| connected            | Connection established              |\n| message              | General message event               |\n| error                | An error occurs                     |\n| log                  | A log message is emitted            |\n| log-event            | A structured log event is emitted   |\n\nTypedEventEmitter Interface:","hierarchy":{"lvl0":"Sdk","lvl1":"API Reference","lvl2":"Other Events","lvl3":""}},
-{"objectID":"7ff1b31f163c4928ce88a79da3c279d1c1c6351ccf3cc5ad2464d2a03174f308","title":"Related Features","url":"/docs/sdk/api-reference#related-features","content":"Human-in-the-Loop (HITL) - Mark tools with requiresConfirmation: true\nGuardrails Middleware - Enable with middleware: { preset: 'security' }\nConversation History - Use getConversationHistory() method\nMultimodal Chat - Use images array in generate() options\nAuto Evaluation - Enable with enableEvaluation: true\nCLI Loop Sessions - Interactive mode with persistent state\nProvider Orchestration - Set enableOrchestration: true\nRegional Streaming - Use region parameter in generate()\nOffice Documents - Use officeFiles array for DOCX, PPTX, XLSX\nPDF Support - Use pdfFiles array for PDF documents\nCSV Support - Use csvFiles array for spreadsheet data\nCLI Commands Reference - CLI equivalents for all SDK methods\nConfiguration Guide - Environment variables and config files\nTroubleshooting - Common SDK issues and solutions\n\nBack to Main README | Next: Visual Demos","hierarchy":{"lvl0":"Sdk","lvl1":"API Reference","lvl2":"Related Features","lvl3":""}},
+{"objectID":"7ff1b31f163c4928ce88a79da3c279d1c1c6351ccf3cc5ad2464d2a03174f308","title":"Related Features","url":"/docs/sdk/api-reference#related-features","content":"Human-in-the-Loop (HITL) - Mark tools with requiresConfirmation: true\nGuardrails Middleware - Set middleware per call, in new NeuroLink({ middleware }) or with configureGlobalMiddleware(); add badWords or modelFilter to filter\nConversation History - Use getConversationHistory() method\nMultimodal Chat - Use images array in generate() options\nAuto Evaluation - Enable with enableEvaluation: true\nCLI Loop Sessions - Interactive mode with persistent state\nProvider Orchestration - Set enableOrchestration: true\nRegional Streaming - Use region parameter in generate()\nOffice Documents - Use officeFiles array for DOCX, PPTX, XLSX\nPDF Support - Use pdfFiles array for PDF documents\nCSV Support - Use csvFiles array for spreadsheet data\nCLI Commands Reference - CLI equivalents for all SDK methods\nConfiguration Guide - Environment variables and config files\nTroubleshooting - Common SDK issues and solutions\n\nBack to Main README | Next: Visual Demos","hierarchy":{"lvl0":"Sdk","lvl1":"API Reference","lvl2":"Related Features","lvl3":""}},
 {"objectID":"72ff7213f71cd3e590c196c5a1a39a7ab2af152776936e1a647f228cae34de6e","title":"🔧 SDK Custom Tools Guide","url":"/docs/sdk/custom-tools-guide","content":"🔧 SDK Custom Tools Guide\n\nBuild powerful AI applications by extending NeuroLink with your own custom tools.\n\n📋 Overview\n\nNeuroLink's SDK allows you to register custom tools programmatically, giving your AI assistants access to any functionality you need. All registered tools work seamlessly with the built-in tool system across all supported providers.\n\nKey Features\n✅ Type-Safe: Full TypeScript support with Zod schema validation\n✅ Provider Agnostic: Works with all providers that support tools\n✅ Easy Integration: Simple API for tool registration\n✅ Async Support: All tools run asynchronously\n✅ Error Handling: Graceful error handling built-in\n\n🚀 Quick Start\n\nBasic Tool Registration\n\n⚠️ Common Mistakes\n\n❌ Using schema instead of parameters\n\n❌ Using plain JSON schema as parameters\n\n✅ Correct Zod Schema Format\n\n📖 SimpleTool Interface\n\nAll custom tools implement the SimpleTool interface:\n\nInterface Components\ndescription: Clear, actionable description that helps the AI understand when to use the tool\nparameters: Optional Zod schema for validating inputs (highly recommended)\nexecute: Async function that implements the tool's logic\n\n🛠️ Registration Methods\n\nRegister Single Tool\n\nRegister Multiple Tools\n\nGet Custom Tools\n\n💡 Common Use Cases\nAPI Integration\nDatabase Operations\nData Processing\nFile Operations\nExternal Service Integration\n\n🎯 Best Practices\nClear Descriptions\n\nMake tool descriptions specific and actionable:\nParameter Validation\n\nAlways use Zod schemas for type safety:\nError Handling\n\nHandle errors gracefully:\n\nA returned error is classified as a failure everywhere the SDK reports one,\nexactly as a thrown error is. A result is error-shaped when it carries\nisError: true (the MCP convention), a non-null error field ({ error:\n\"User not found\" }), or a status of \"error\" / \"failed\" / \"failure\" /\n\"fail\". Such a call emits tool:end with success: false and the error\ntext, is stored in session memory as success: false, replays on later turns\nas [called <tool> → error], and is reported with isError: true in\nresult.toolExecutions. The model still receives the full object you\nreturned — classification only affects telemetry and memory, never what the\nmodel sees.\nAsync Operations\n\nAll execute functions must return promises:\nTool Naming\n\nUse clear, consistent naming:\n\n🧪 Testing Your Tools\n\nUnit Testing\n\nIntegration Testing\n\n🔍 Debugging Tools\n\nEnable Debug Mode\n\nLog Tool Execution\n\n🚀 Advanced Patterns\n\nTool Composition\n\nTool Middleware\n\nDynamic Tool Registration\n\nForcing a Tool Call\n\ntoolChoice tells the model whether it must call a tool on a step:\n\n| Value                             | Effect                                           |\n| --------------------------------- | ------------------------------------------------ |\n| \"auto\" (default)                | The model decides whether and which tool to call |\n| \"none\"                          | No tool calls on any step                        |\n| \"required\"                      | The model must call at least one tool            |\n| { type: \"tool\", toolName: \"x\" } | The model must call the named tool               |\n\nA forced choice (\"required\" or a named tool) is applied only to the first\ntoolChoiceSteps steps of the tool loop — default 1 — after which the\nmodel is free to answer. Holding a forced choice on every step would compel\na tool call on every step, and the turn could only end when maxSteps ran\nout. \"auto\" and \"none\" are applied unchanged on every step.\n\nFor anything finer, prepareStep runs before every step and may return a\ntoolChoice for that step, which takes precedence over toolChoiceSteps:\n\nprepareStep receives { stepNumber, steps, maxSteps, model }, where steps\nholds one record per completed step (its text, tool calls — with input\nalready parsed into an object — tool results, finish reason and usage,\nincluding prompt-cache tokens where the provider reports them). Only the\nreturned toolChoice is honoured; the legacy model and\nexperimental_activeTools / activeTools fields are accepted for source\ncompatibility but ignored, and their presence is logged once as a warning —\nuse toolFilter / excludeTools to scope tools. A hook that throws is\nlogged and that step falls back to the toolChoice / toolChoiceSteps rule;\na hook still pending when the turn is aborted or times out is released with\nthe abort. Both toolChoice and prepareStep apply to stream() as well.\n\nVertex, Google AI Studio and Bedrock do not honour toolChoice at all today.\nSageMaker honours tools, toolChoice and prepareStep on generate()\nonly: its stream() sends the prompt without tools.\n\n📊 Performance Considerations\nTimeout Handling\nCaching\nBatch Operations\n\n🔒 Security Considerations\n\nInput Sanitization\n\nPermission Checking\n\nRate Limiting\n\n🎉 Complete Example\n\nHere's a complete example combining multiple concepts:\n\n🌐 MCP Server Integration\n\nBeyond simple tool registration, NeuroLink SDK supports adding complete MCP (Model Context Protocol) servers for more complex tool ecosystems.\n\nAdding I","hierarchy":{"lvl0":"Sdk","lvl1":"🔧 SDK Custom Tools Guide","lvl2":"","lvl3":""}},
 {"objectID":"fe570104af53b04cac97749a24ce61018428abe7df180a14a7285ebf6fda6fca","title":"🔧 SDK Custom Tools Guide","url":"/docs/sdk/custom-tools-guide#-sdk-custom-tools-guide","content":"Build powerful AI applications by extending NeuroLink with your own custom tools.","hierarchy":{"lvl0":"Sdk","lvl1":"🔧 SDK Custom Tools Guide","lvl2":"🔧 SDK Custom Tools Guide","lvl3":""}},
 {"objectID":"4c744b8358a0761297815b9e1244bad43bebb4ece00ddac4cb24b186dbae5647","title":"📋 Overview","url":"/docs/sdk/custom-tools-guide#-overview","content":"NeuroLink's SDK allows you to register custom tools programmatically, giving your AI assistants access to any functionality you need. All registered tools work seamlessly with the built-in tool system across all supported providers.","hierarchy":{"lvl0":"Sdk","lvl1":"🔧 SDK Custom Tools Guide","lvl2":"📋 Overview","lvl3":""}},
@@ -11186,8 +11214,8 @@
 {"objectID":"1a0933649d2e2a9e068e520585c510992a787eb886f8661e6da36cf83ce7e3eb","title":"Multiple inputs","url":"/docs/skills/neurolink-guide/multimodal#multiple-inputs","content":"neurolink generate \"Compare\" --image ./a.png --image ./b.png --pdf ./docs.pdf\n`","hierarchy":{"lvl0":"Skills","lvl1":"NeuroLink Multimodal Support","lvl2":"Multiple inputs","lvl3":""}},
 {"objectID":"e2858a465b5d2bc8ef1db9834500597eb845bf3b40f3a0ba4b29e54033a4b23b","title":"File Size Considerations","url":"/docs/skills/neurolink-guide/multimodal#file-size-considerations","content":"| File Type | Recommended Max | Notes                |\n| --------- | --------------- | -------------------- |\n| Images    | 20MB            | Resized if larger    |\n| PDFs      | 50MB            | Page limit may apply |\n| CSV       | 10MB            | Use maxRows option   |\n| Code      | 100KB           | Split large files    |","hierarchy":{"lvl0":"Skills","lvl1":"NeuroLink Multimodal Support","lvl2":"File Size Considerations","lvl3":""}},
 {"objectID":"8d5a67779513f44569761e7ea2fa399de1aafd682a08e0b14c08850f0c223cb2","title":"Next Steps","url":"/docs/skills/neurolink-guide/multimodal#next-steps","content":"MCP tools - Add external tools\nRAG integration - Document-grounded generation\nProviders - Configure vision-capable providers","hierarchy":{"lvl0":"Skills","lvl1":"NeuroLink Multimodal Support","lvl2":"Next Steps","lvl3":""}},
-{"objectID":"d9e19daaa21e9512ad71f1187055c474d4b51da0c78b56a197286c6403126b9a","title":"NeuroLink Provider Configuration","url":"/docs/skills/neurolink-guide/providers","content":"NeuroLink Provider Configuration\n\nNeuroLink supports many AI providers through a unified API. This page highlights the most commonly-configured text providers — see the README provider table and the Provider Capabilities Audit for the full matrix, including newer text providers (DeepSeek, NVIDIA NIM, LM Studio, llama.cpp, plus the Tier-2 catalog providers — Groq, Cerebras, SambaNova, Together AI, Fireworks AI, Perplexity, Cloudflare Workers AI, xAI, and more), the decision-only TypeSafe Jev, Laya, XOR, SubQ, Perplexity Decisions, Cloudflare Clef and OpenAI Decisions providers (serve decide(), not generate()/stream(); tried in the order TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef, OpenAI Decisions when several are configured; Clef reads the same CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID as the Cloudflare Workers AI text provider, and OpenAI Decisions the same OPENAI_API_KEY as the OpenAI chat provider), and voice providers (OpenAI TTS, ElevenLabs, Deepgram, Azure Speech, Whisper, Fish Audio, Cartesia, OpenAI Realtime, Gemini Live).\n\nCommon Providers\n\n| Provider         | Enum Name      | Aliases        | Default Model                           |\n| ---------------- | -------------- | -------------- | --------------------------------------- |\n| OpenAI           | openai       | gpt, chatgpt   | gpt-4o                                  |\n| Anthropic        | anthropic    | claude         | claude-3-5-sonnet-20241022              |\n| Google AI Studio | google-ai    | gemini, google | gemini-2.5-flash                        |\n| Google Vertex AI | vertex       | google-vertex  | gemini-2.5-flash                        |\n| AWS Bedrock      | bedrock      | aws-bedrock    | anthropic.claude-3-sonnet-20240229-v1:0 |\n| Azure OpenAI     | azure-openai | azure          | gpt-4o                                  |\n| Mistral AI       | mistral      | -              | mistral-large                           |\n| Ollama           | ollama       | -              | llama3                                  |\n| LiteLLM          | litellm      | -              | varies                                  |\n| AWS SageMaker    | sagemaker    | -              | custom                                  |\n| Hugging Face     | hugging-face | hf             | varies                                  |\n| OpenRouter       | openrouter   | -              | varies                                  |\n| Gateway          | gateway      | -              | varies                                  |\n\nOpenAI\n\nAvailable Models:\ngpt-4o - Latest GPT-4 Omni\ngpt-4o-mini - Faster, cheaper\ngpt-4-turbo - GPT-4 Turbo\no1 - Reasoning model\no1-mini - Smaller reasoning model\n\nAnthropic\n\nAvailable Models:\nclaude-3-5-sonnet-20241022 - Latest Sonnet\nclaude-3-7-sonnet-20250219 - Claude 3.7 Sonnet\nclaude-3-opus-20240229 - Most capable\nclaude-3-haiku-20240307 - Fastest\n\nExtended Thinking:\n\nGoogle AI Studio\n\nAvailable Models:\ngemini-2.5-flash - Fast and capable\ngemini-2.5-pro - Most capable\ngemini-2.0-flash - Previous generation\ngemini-3-flash-preview - Preview of Gemini 3\n\nGoogle Vertex AI\n\nAvailable Models:\ngemini-3-flash - Latest Gemini 3\ngemini-3-pro - Most capable Gemini 3\ngemini-2.5-flash - Fast\ngemini-2.5-pro - Previous gen capable\n\nExtended Thinking (Gemini 3):\n\nAWS Bedrock\n\nAvailable Models:\nanthropic.claude-3-sonnet-20240229-v1:0\nanthropic.claude-3-haiku-20240307-v1:0\nanthropic.claude-3-opus-20240229-v1:0\namazon.titan-text-express-v1\namazon.nova-pro-v1:0\nmeta.llama3-70b-instruct-v1:0\n\nAzure OpenAI\n\nMistral AI\n\nAvailable Models:\nmistral-large-latest - Most capable\nmistral-small-latest - Fast\ncodestral-latest - Code specialized\nministral-8b-latest - Small\n\nOllama (Local)\n\nSetup:\n\nAvailable Models:\nllama3 - Meta Llama 3\nllama3:70b - Larger Llama 3\nmistral - Mistral 7B\ncodellama - Code specialized\nphi3 - Microsoft Phi-3\n\nLiteLLM\n\nAWS SageMaker\n\nHugging Face\n\nOpenRouter\n\nProvider Fallback\n\nConfigure automatic fallback to another provider:\n\nCheck Provider Status\n\nProvider-Specific Options\n\nTemperature and Sampling\n\nSystem Prompts\n\nVision-Capable Models\n\nNot all models support image inputs:\n\n| Provider  | Vision Models         |\n| --------- | --------------------- |\n| OpenAI    | gpt-4o, gpt-4-turbo   |\n| Anthropic | All Claude 3 models   |\n| Vertex    | Gemini 2.5+, Gemini 3 |\n| Google AI | Gemini 2.5+, Gemini 3 |\n| Bedrock   | Claude 3 models       |\n\nNext Steps\nMultimodal inputs - Work with images and documents\nMCP tools - Add external tools\nRAG integration - Document-grounded generation","hierarchy":{"lvl0":"Skills","lvl1":"NeuroLink Provider Configuration","lvl2":"","lvl3":""}},
-{"objectID":"b66a3f5004ec2abbc42277973845a5485480e16a3ea54ea9f31e83978a9ee521","title":"NeuroLink Provider Configuration","url":"/docs/skills/neurolink-guide/providers#neurolink-provider-configuration","content":"NeuroLink supports many AI providers through a unified API. This page highlights the most commonly-configured text providers — see the README provider table and the Provider Capabilities Audit for the full matrix, including newer text providers (DeepSeek, NVIDIA NIM, LM Studio, llama.cpp, plus the Tier-2 catalog providers — Groq, Cerebras, SambaNova, Together AI, Fireworks AI, Perplexity, Cloudflare Workers AI, xAI, and more), the decision-only TypeSafe Jev, Laya, XOR, SubQ, Perplexity Decisions, Cloudflare Clef and OpenAI Decisions providers (serve decide(), not generate()/stream(); tried in the order TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef, OpenAI Decisions when several are configured; Clef reads the same CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID as the Cloudflare Workers AI text provider, and OpenAI Decisions the same OPENAI_API_KEY as the OpenAI chat provider), and voice providers (OpenAI TTS, ElevenLabs, Deepgram, Azure Speech, Whisper, Fish Audio, Cartesia, OpenAI Realtime, Gemini Live).","hierarchy":{"lvl0":"Skills","lvl1":"NeuroLink Provider Configuration","lvl2":"NeuroLink Provider Configuration","lvl3":""}},
+{"objectID":"d9e19daaa21e9512ad71f1187055c474d4b51da0c78b56a197286c6403126b9a","title":"NeuroLink Provider Configuration","url":"/docs/skills/neurolink-guide/providers","content":"NeuroLink Provider Configuration\n\nNeuroLink supports many AI providers through a unified API. This page highlights the most commonly-configured text providers — see the README provider table and the Provider Capabilities Audit for the full matrix, including newer text providers (DeepSeek, NVIDIA NIM, LM Studio, llama.cpp, plus the Tier-2 catalog providers — Groq, Cerebras, SambaNova, Together AI, Fireworks AI, Perplexity, Cloudflare Workers AI, xAI, and more), the decision-only TypeSafe Jev, Laya, XOR, SubQ, Perplexity Decisions, Cloudflare Clef, OpenAI Decisions and ScaleDown providers (serve decide(), not generate()/stream(); tried in the order TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef, OpenAI Decisions, ScaleDown when several are configured; Clef reads the same CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID as the Cloudflare Workers AI text provider, and OpenAI Decisions the same OPENAI_API_KEY as the OpenAI chat provider; ScaleDown has its own SCALEDOWN_API_KEY), and voice providers (OpenAI TTS, ElevenLabs, Deepgram, Azure Speech, Whisper, Fish Audio, Cartesia, OpenAI Realtime, Gemini Live).\n\nCommon Providers\n\n| Provider         | Enum Name      | Aliases        | Default Model                           |\n| ---------------- | -------------- | -------------- | --------------------------------------- |\n| OpenAI           | openai       | gpt, chatgpt   | gpt-4o                                  |\n| Anthropic        | anthropic    | claude         | claude-3-5-sonnet-20241022              |\n| Google AI Studio | google-ai    | gemini, google | gemini-2.5-flash                        |\n| Google Vertex AI | vertex       | google-vertex  | gemini-2.5-flash                        |\n| AWS Bedrock      | bedrock      | aws-bedrock    | anthropic.claude-3-sonnet-20240229-v1:0 |\n| Azure OpenAI     | azure-openai | azure          | gpt-4o                                  |\n| Mistral AI       | mistral      | -              | mistral-large                           |\n| Ollama           | ollama       | -              | llama3                                  |\n| LiteLLM          | litellm      | -              | varies                                  |\n| AWS SageMaker    | sagemaker    | -              | custom                                  |\n| Hugging Face     | hugging-face | hf             | varies                                  |\n| OpenRouter       | openrouter   | -              | varies                                  |\n| Gateway          | gateway      | -              | varies                                  |\n\nOpenAI\n\nAvailable Models:\ngpt-4o - Latest GPT-4 Omni\ngpt-4o-mini - Faster, cheaper\ngpt-4-turbo - GPT-4 Turbo\no1 - Reasoning model\no1-mini - Smaller reasoning model\n\nAnthropic\n\nAvailable Models:\nclaude-3-5-sonnet-20241022 - Latest Sonnet\nclaude-3-7-sonnet-20250219 - Claude 3.7 Sonnet\nclaude-3-opus-20240229 - Most capable\nclaude-3-haiku-20240307 - Fastest\n\nExtended Thinking:\n\nGoogle AI Studio\n\nAvailable Models:\ngemini-2.5-flash - Fast and capable\ngemini-2.5-pro - Most capable\ngemini-2.0-flash - Previous generation\ngemini-3-flash-preview - Preview of Gemini 3\n\nGoogle Vertex AI\n\nAvailable Models:\ngemini-3-flash - Latest Gemini 3\ngemini-3-pro - Most capable Gemini 3\ngemini-2.5-flash - Fast\ngemini-2.5-pro - Previous gen capable\n\nExtended Thinking (Gemini 3):\n\nAWS Bedrock\n\nAvailable Models:\nanthropic.claude-3-sonnet-20240229-v1:0\nanthropic.claude-3-haiku-20240307-v1:0\nanthropic.claude-3-opus-20240229-v1:0\namazon.titan-text-express-v1\namazon.nova-pro-v1:0\nmeta.llama3-70b-instruct-v1:0\n\nAzure OpenAI\n\nMistral AI\n\nAvailable Models:\nmistral-large-latest - Most capable\nmistral-small-latest - Fast\ncodestral-latest - Code specialized\nministral-8b-latest - Small\n\nOllama (Local)\n\nSetup:\n\nAvailable Models:\nllama3 - Meta Llama 3\nllama3:70b - Larger Llama 3\nmistral - Mistral 7B\ncodellama - Code specialized\nphi3 - Microsoft Phi-3\n\nLiteLLM\n\nAWS SageMaker\n\nHugging Face\n\nOpenRouter\n\nProvider Fallback\n\nConfigure automatic fallback to another provider:\n\nCheck Provider Status\n\nProvider-Specific Options\n\nTemperature and Sampling\n\nSystem Prompts\n\nVision-Capable Models\n\nNot all models support image inputs:\n\n| Provider  | Vision Models         |\n| --------- | --------------------- |\n| OpenAI    | gpt-4o, gpt-4-turbo   |\n| Anthropic | All Claude 3 models   |\n| Vertex    | Gemini 2.5+, Gemini 3 |\n| Google AI | Gemini 2.5+, Gemini 3 |\n| Bedrock   | Claude 3 models       |\n\nNext Steps\nMultimodal inputs - Work with images and documents\nMCP tools - Add external tools\nRAG integration - Document-grounded generation","hierarchy":{"lvl0":"Skills","lvl1":"NeuroLink Provider Configuration","lvl2":"","lvl3":""}},
+{"objectID":"b66a3f5004ec2abbc42277973845a5485480e16a3ea54ea9f31e83978a9ee521","title":"NeuroLink Provider Configuration","url":"/docs/skills/neurolink-guide/providers#neurolink-provider-configuration","content":"NeuroLink supports many AI providers through a unified API. This page highlights the most commonly-configured text providers — see the README provider table and the Provider Capabilities Audit for the full matrix, including newer text providers (DeepSeek, NVIDIA NIM, LM Studio, llama.cpp, plus the Tier-2 catalog providers — Groq, Cerebras, SambaNova, Together AI, Fireworks AI, Perplexity, Cloudflare Workers AI, xAI, and more), the decision-only TypeSafe Jev, Laya, XOR, SubQ, Perplexity Decisions, Cloudflare Clef, OpenAI Decisions and ScaleDown providers (serve decide(), not generate()/stream(); tried in the order TypeSafe, Laya, XOR, SubQ, Perplexity, Cloudflare Clef, OpenAI Decisions, ScaleDown when several are configured; Clef reads the same CLOUDFLARE_API_KEY and CLOUDFLARE_ACCOUNT_ID as the Cloudflare Workers AI text provider, and OpenAI Decisions the same OPENAI_API_KEY as the OpenAI chat provider; ScaleDown has its own SCALEDOWN_API_KEY), and voice providers (OpenAI TTS, ElevenLabs, Deepgram, Azure Speech, Whisper, Fish Audio, Cartesia, OpenAI Realtime, Gemini Live).","hierarchy":{"lvl0":"Skills","lvl1":"NeuroLink Provider Configuration","lvl2":"NeuroLink Provider Configuration","lvl3":""}},
 {"objectID":"ca77e6781279a1e849f17c5edd76b2cc111546faa201202d3dc11820334170ce","title":"Common Providers","url":"/docs/skills/neurolink-guide/providers#common-providers","content":"| Provider         | Enum Name      | Aliases        | Default Model                           |\n| ---------------- | -------------- | -------------- | --------------------------------------- |\n| OpenAI           | openai       | gpt, chatgpt   | gpt-4o                                  |\n| Anthropic        | anthropic    | claude         | claude-3-5-sonnet-20241022              |\n| Google AI Studio | google-ai    | gemini, google | gemini-2.5-flash                        |\n| Google Vertex AI | vertex       | google-vertex  | gemini-2.5-flash                        |\n| AWS Bedrock      | bedrock      | aws-bedrock    | anthropic.claude-3-sonnet-20240229-v1:0 |\n| Azure OpenAI     | azure-openai | azure          | gpt-4o                                  |\n| Mistral AI       | mistral      | -              | mistral-large                           |\n| Ollama           | ollama       | -              | llama3                                  |\n| LiteLLM          | litellm      | -              | varies                                  |\n| AWS SageMaker    | sagemaker    | -              | custom                                  |\n| Hugging Face     | hugging-face | hf             | varies                                  |\n| OpenRouter       | openrouter   | -              | varies                                  |\n| Gateway          | gateway      | -              | varies                                  |","hierarchy":{"lvl0":"Skills","lvl1":"NeuroLink Provider Configuration","lvl2":"Common Providers","lvl3":""}},
 {"objectID":"9c5370e9c625cd036ee1c3c8ba7c0bea9a5b6b68ca08ccd9c710d3e65429d843","title":"OpenAI","url":"/docs/skills/neurolink-guide/providers#openai","content":"`bash","hierarchy":{"lvl0":"Skills","lvl1":"NeuroLink Provider Configuration","lvl2":"OpenAI","lvl3":""}},
 {"objectID":"fcc21bd062d4e7ab394bd0ad8a1257ec9ed367c205f68dab30479e6758c5cf20","title":"Environment","url":"/docs/skills/neurolink-guide/providers#environment","content":"OPENAIAPIKEY=sk-...\nOPENAIORGID=org-...           # Optional\nOPENAIBASEURL=...             # Optional, for proxies\ntypescript\nconst result = await neurolink.generate({\n  input: { text: \"Hello\" },\n  provider: \"openai\",\n  model: \"gpt-4o\", // or gpt-4o-mini, gpt-4-turbo, o1, o1-mini\n});\n\n\n**Available Models:**\n\n- gpt-4o - Latest GPT-4 Omni\n- gpt-4o-mini - Faster, cheaper\n- gpt-4-turbo - GPT-4 Turbo\n- o1 - Reasoning model\n- o1-mini` - Smaller reasoning model","hierarchy":{"lvl0":"Skills","lvl1":"NeuroLink Provider Configuration","lvl2":"Environment","lvl3":""}},