chat-agent-toolkit 1.2.1

package/README.md

@@ -0,0 +1,234 @@
+# agent-toolkit
+
+
+Multi-provider AI agent toolkit for generating language responses, searching the web, extracting page content, and managing long-term memory across 10+ LLM providers.
+
+Built on top of the [Vercel AI SDK](https://sdk.vercel.ai), with a small registry of pre-tuned agent prompts (research, summarization, citation answering, query resolution, knowledge-graph extraction, etc.) and tool wrappers around the [QwkSearch](https://qwksearch.com) API.
+
+## Language Intelligence Providers
+
+| Provider               | 🌍  | Top Model (Others)                            | 🏆 Benchmarks                                                         | 📄 Docs                                                                                               | 🔑 Keys                                                                                            | 💰 Funding   |
+| ---------------------- | --- | --------------------------------------------- | --------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------- | ------------ |
+| **Anthropic**          | 🇺🇸  | Claude Mythos / Opus  (Sonnet, Haiku)         | 🥇 GPQA Diamond 94.6% · 🥇 SWE-bench 93.9% · 🧬 PhD reasoning         | [Docs](https://docs.anthropic.com/en/docs/welcome)                                                    | [Keys](https://console.anthropic.com/settings/keys)                                                | ~$60B        |
+| **OpenAI**             | 🇺🇸  | GPT / o3 / Codex (o1, o4, o4-mini, gpt-4o)    | 🥇 AIME 2025 100% · 🥇 SWE-bench Pro · 📚 MMLU-Pro 90%                | [Docs](https://platform.openai.com/docs/overview)                                                     | [Keys](https://platform.openai.com/api-keys)                                                       | ~$180B       |
+| **Google**      | 🇺🇸  | Gemini Pro (Flash, Flash-Lite, Gemma)         | 🥇 GPQA 94.1% · 🥇 LiveCodeBench Elo 2439 · 🌐 #1 in 6/13 Vals        | [Docs](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models)                            | [Keys](https://cloud.google.com/vertex-ai/generative-ai/docs/start/express-mode/overview#api-keys) | Public       |
+| **xAI**                | 🇺🇸  | Grok Heavy (Grok-3, Grok Vision)              | 🥇 AIME 2025 100% · 🧮 Math competition · ⚡ X integration            | [Docs](https://docs.x.ai/docs#models)                                                                 | [Keys](https://console.x.ai/)                                                                      | ~$45B        |
+| **Meta**               | 🇺🇸  | Llama Maverick / Scout (Llama 3.x, CodeLlama) | 🥇 DocVQA 94.4% · 🥇 10M token context · 📊 ChartQA 90%               | [Docs](https://www.llama.com/docs/overview/)                                                          | [Keys](https://www.llama.com/llama-downloads/)                                                     | Public       |
+| **NVIDIA**             | 🇺🇸  | Nemotron-Cascade  (Llama Nemotron, Kimi)      | 🥇 LCB v6 87.2% · 🏅 IMO+IOI+ICPC gold · 🧮 AIME 98.6%                | [Docs](https://docs.api.nvidia.com/nim/reference/llm-apis)                                            | [Keys](https://build.nvidia.com/settings/api-keys)                                                 | Public       |
+| **Perplexity**      | 🇺🇸  | Sonar Reasoning Pro (Sonar Deep Research)     | 🥇 Search Arena · 🔍 #1 web-grounded QA · 🌐 Real-time retrieval      | [Docs](https://docs.perplexity.ai/models/model-cards)                                                 | [Keys](https://www.perplexity.ai/account/api/keys)                                                 | ~$1B         |
+| **Groq**               | 🇺🇸  | (Llama, DeepSeek, Gemma, Mistral, Qwen)       | ⚡ #1 inference speed · 🏎️ Fastest TTFT · 🔧 LPU hardware             | [Docs](https://console.groq.com/docs/overview)                                                        | [Keys](https://console.groq.com/keys)                                                              | ~$640M       |
+| **Mistral**         | 🇫🇷  | Mistral Large  (Small 4, Codestral, Devstral) | 🥈 Arena Elo 1418 · 🌍 Multilingual MMLU 85.5% · 🚀 Fastest TTFT      | [Docs](https://docs.mistral.ai/)                                                                      | [Keys](https://console.mistral.ai/api-keys/)                                                       | ~$3.1B       |
+| **Together**        | 🇺🇸  | (Llama, Mistral, Gemma, Qwen, DeepSeek)       | 🏗️ Widest open hosting · 💸 Best open-source pricing · 🔧 Fine-tuning | [Docs](https://docs.together.ai/docs/quickstart)                                                      | [Keys](https://api.together.xyz/settings/api-keys)                                                 | ~$225M       |
+| **Moonshot** | 🇨🇳  | Kimi Reasoning (K2.6, K2)                     | 🥇 AIME open 96.1% · 🥇 MATH-500 98% · 🥇 HumanEval 99%               | [Docs](https://platform.moonshot.cn/docs)                                                             | [Keys](https://platform.moonshot.cn/console/api-keys)                                              | ~$3.9B       |
+| **Zhipu**     | 🇨🇳  | GLM Reasoning / GLM-4.7 (GLM-4V, CogView)     | 🥇 Chatbot Arena Elo 1451 · 🥇 MMLU 96% · 🧮 AIME 95.7%               | [Docs](https://bigmodel.cn/dev/api)                                                                   | [Keys](https://bigmodel.cn/usercenter/apikeys)                                                     | ~$1.8B       |
+| **Alibaba**     | 🇨🇳  | Qwen-Coder / Qwen  (Qwen-VL, Qwen-Audio)      | 🥇 Codeforces Elo 2056 · 💻 SWE-bench 69.6% · 🏎️ LCB 70.7%            | [Docs](https://www.alibabacloud.com/help/en/model-studio/developer-reference/use-qwen-by-calling-api) | [Keys](https://bailian.console.aliyun.com/?apiKey=1)                                               | Public       |
+| **DeepSeek**           | 🇨🇳  | DeepSeek (DeepSeek-Coder, DeepSeek-VL)        | 🥇 IMO gold (open) · 📚 MMLU-Pro 81.2 · 🧮 AIME 87.5%                 | [Docs](https://api-docs.deepseek.com/)                                                                | [Keys](https://platform.deepseek.com/api_keys)                                                     | Bootstrapped |
+| **Cloudflare**         | 🇺🇸  | (Llama, Mistral, Gemma, Qwen, DeepSeek)       | 🌐 Edge inference · ⚡ Serverless CDN scale · 🔒 Privacy-first        | [Docs](https://developers.cloudflare.com/workers-ai/)                                                 | [Keys](https://dash.cloudflare.com/profile/api-tokens)                                             | Public       |
+| **Ollama**             | 🇺🇸  | (Llama, Mistral, Gemma, Qwen, DeepSeek)       | 🖥️ #1 local inference · 🔒 Fully offline · 🆓 Free self-hosted        | [Docs](https://ollama.com/docs)                                                                       | [Keys](https://ollama.com/settings/keys)                                                           | ~$20M        |
+
+## Amazon Bedrock Provider
+
+Use Vercel AI SDK's official Bedrock provider: install `@ai-sdk/amazon-bedrock`, configure AWS auth, then call `bedrock('model-id')` in `generateText` or `streamText`. [ai-sdk](https://ai-sdk.dev/providers/ai-sdk-providers/amazon-bedrock)
+
+### Install
+
+- Run `pnpm add ai @ai-sdk/amazon-bedrock` or the npm/yarn equivalent. [vercel](https://vercel.com/changelog/amazon-bedrock-provider-for-the-vercel-ai-sdk-now-available)
+- Import either `bedrock` directly or `createAmazonBedrock` for custom config. [ai-sdk](https://ai-sdk.dev/providers/ai-sdk-providers/amazon-bedrock)
+
+### Env setup
+
+- Simplest path: set `AWS_BEARER_TOKEN_BEDROCK` with a Bedrock API key; the docs say API key auth is the recommended method. [ai-sdk](https://ai-sdk.dev/providers/ai-sdk-providers/amazon-bedrock)
+- SigV4 also works with `AWS_ACCESS_KEY_ID`, `AWS_SECRET_ACCESS_KEY`, and `AWS_REGION`, and the provider can also use the AWS credential chain. [ai-sdk](https://ai-sdk.dev/providers/ai-sdk-providers/amazon-bedrock)
+
+### Minimal example
+
+```ts
+import { generateText } from 'ai';
+import { bedrock } from '@ai-sdk/amazon-bedrock';
+
+const { text } = await generateText({
+  model: bedrock('anthropic.claude-3-haiku-20240307-v1:0'),
+  prompt: 'Explain Bedrock in one sentence.',
+});
+
+console.log(text);
+```
+
+### Custom region
+
+```ts
+import { createAmazonBedrock } from '@ai-sdk/amazon-bedrock';
+
+export const amazonBedrock = createAmazonBedrock({
+  region: 'us-east-1',
+  apiKey: process.env.AWS_BEARER_TOKEN_BEDROCK,
+});
+```
+
+### Notes
+
+- You must enable model access in the AWS Bedrock console first; access is not granted by default. [ai-sdk](https://ai-sdk.dev/providers/ai-sdk-providers/amazon-bedrock)
+- `streamText` is supported too, and Bedrock-specific options like guardrails can be passed via `providerOptions.bedrock`. [ai-sdk](https://ai-sdk.dev/providers/ai-sdk-providers/amazon-bedrock)
+
+## Install
+
+```bash
+npm install ai-research-agent
+```
+
+## Quick start
+
+```ts
+import { generateLanguageResponse } from "ai-research-agent";
+
+const response = await generateLanguageResponse({
+  provider: "groq",
+  apiKey: process.env.GROQ_API_KEY,
+  agent: "question",
+  query: "Explain transformer attention in two sentences.",
+});
+
+console.log(response.content);
+```
+
+`response.content` is HTML by default (set `html: false` for raw Markdown). Agents that define an `after` hook also return parsed data on `response.extract`.
+
+### Using Amazon Bedrock
+
+```ts
+import { generateLanguageResponse } from "ai-research-agent";
+
+const response = await generateLanguageResponse({
+  provider: "amazon",
+  apiKey: process.env.AWS_BEARER_TOKEN_BEDROCK, // or "region:accessKeyId:secretAccessKey"
+  model: "anthropic.claude-3-5-sonnet-20241022-v2:0",
+  agent: "question",
+  query: "Explain transformer attention in two sentences.",
+});
+
+console.log(response.content);
+```
+
+## Supported providers
+
+| Provider        | ID           | Default model                                   | Notes                                             |
+| --------------- | ------------ | ----------------------------------------------- | ------------------------------------------------- |
+| Anthropic       | `anthropic`  | `claude-3-7-sonnet-20250219`                    |                                                   |
+| OpenAI          | `openai`     | `gpt-4o`                                        |                                                   |
+| Groq            | `groq`       | `meta-llama/llama-4-maverick-17b-128e-instruct` |                                                   |
+| Google Vertex   | `google`     | Gemini 2.x                                      |                                                   |
+| XAI             | `xai`        | `grok-beta`                                     |                                                   |
+| Amazon Bedrock  | `amazon`     | `anthropic.claude-3-5-sonnet-20241022-v2:0`     | apiKey = bearer token or `region:key:secret`     |
+| Cloudflare      | `cloudflare` | `llama-4-scout-17b-16e-instruct`                | apiKey =`token:accountId`                         |
+| Together AI     | `togetherai` | `meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo`   |                                                   |
+| Perplexity      | `perplexity` | `sonar`                                         |                                                   |
+| NVIDIA NIM      | `nvidia`     | `moonshotai/kimi-k2.5`                          |                                                   |
+| Ollama          | `ollama`     | `llama3.2`                                      | Local — no key required                           |
+
+The full registry (including context lengths) is exported as `LANGUAGE_MODELS`.
+
+## Built-in agents
+
+The `agent` option selects a prompt template from [`AGENT_PROMPTS`](src/agents/agent-prompts.ts):
+
+- `question` — general-purpose Q&A with chat history.
+- `question-research-engine` — long-form, journalistic, citation-style answer.
+- `query-resolution` — rephrases a follow-up question into standalone search queries.
+- `query-resolution-search` — classifies a query and proposes search categories as JSON.
+- `summarize`, `summarize-bullets`, `summary-longtext` — single-shot and reduce-style summarization.
+- `answer-cite-sources` — answers a query against a `{context}` block with `[n]` citations.
+- `suggest-followups` — emits 4–5 follow-up questions inside `<suggestions>` tags.
+- `remember-user` — extracts factual user memories as structured JSON.
+- `knowledge-graph-nodes` — builds a temporal knowledge graph from a document.
+- `results-relevance-filter` — picks the most relevant URLs from a search result list.
+
+Templates use `{variableName}` placeholders that are filled from the options object (`query`, `article`, `chat_history`, `context`, etc.).
+
+## Agent tools
+
+`AGENT_TOOLS` registers callable tools that the ReAct agent attaches automatically when a prompt declares them in its `tools` array:
+
+- `web_search` — QwkSearch metasearch over 100+ engines.
+- `extract_page` — Mozilla Readability + Mercury extraction with PDF and YouTube support.
+- `generate_ai_response` — proxy to QwkSearch's hosted `/language` endpoint.
+
+Set `QWKSEARCH_URL` and `QWKSEARCH_API_KEY` in the environment, or pass `baseURL`/`apiKey` per-call.
+
+## Memory
+
+```ts
+import {
+  SimpleMemory,
+  MemoryAgent,
+  DrizzleMemoryStorage,
+} from "ai-research-agent";
+```
+
+- `SimpleMemory` — in-memory store implementing `IMemoryStorage`.
+- `DrizzleMemoryStorage` — Drizzle ORM-backed persistence; pair with `createMemorySchema()`.
+- `MemoryAgent` — high-level wrapper that uses the `remember-user` agent to extract facts from chat turns and writes them to a storage backend.
+
+See [src/memory/README.md](src/memory/README.md) and [src/memory/ARCHITECTURE.md](src/memory/ARCHITECTURE.md) for details.
+
+## Package layout
+
+```
+src/
+  agents/                    # prompt + tool registry, generate function, model list
+    agent-prompts.ts         # AGENT_PROMPTS, extractJSONFromLanguageReply
+    agent-tools.ts           # AGENT_TOOLS (web_search, extract_page, ...)
+    generate-language.ts     # generateLanguageResponse — main entry point
+    generate-language-types.ts
+    language-model-names.ts  # LANGUAGE_MODELS, LANGUAGE_PROVIDERS
+    llm-providers.ts         # createLLMProvider — chat-model factory
+    index.ts                 # barrel export
+  memory/                    # SimpleMemory, MemoryAgent, Drizzle storage
+  providers/                 # alternate provider abstractions (model registry)
+  utils/                     # markdown-to-html, document helpers
+  index.ts                   # public entry — re-exports agents + memory
+```
+
+## Build
+
+```bash
+npm run build      # vite build
+npm run make       # clean + build with extra heap
+npm run test       # vitest
+```
+
+Vite bundles ES + CJS targets to `dist/`, emits `.d.ts` files alongside, and applies a `fs/promises` polyfill so the bundle works in browser and edge runtimes (Cloudflare Workers, Vercel Edge).
+
+## Resources
+
+### Documentation & Guides
+- [Vercel AI SDK generateText docs](https://sdk.vercel.ai/docs/reference/ai-sdk-core/generate-text)
+- [Hugging Face tutorials](https://huggingface.co/learn)
+- [Illustrated Transformer](https://jalammar.github.io/illustrated-transformer/)
+- [Building a Transformer with PyTorch](https://www.datacamp.com/tutorial/building-a-transformer-with-py-torch)
+- [LLM training example](https://github.com/vtempest/ai-research-agent/blob/master/packages/neural-net/src/train/predict-next-word.js)
+
+### Transformer Architecture Visualizations
+
+<img src="https://i.imgur.com/uW6E9VJ.gif" alt="Transformer architecture visualization" />
+
+
+## Alternative Agents Frameworks
+
+| # | Framework | Features | Stars | Language | Maker |
+|---:|---|---|---:|---|---|
+| 1 | [Hermes Agent](https://github.com/NousResearch/hermes-agent) | Self-improving agent with a built-in learning loop; distills reusable skills from past runs, persists experience across sessions, "grows with you" over time | 211k | Python | Nous Research |
+| 2 | [AutoGPT](https://github.com/Significant-Gravitas/AutoGPT) | One of the original autonomous agents; goal-driven loop that self-prompts to decompose and execute tasks. Now a low-code platform: visual block-based builder, deploy/run continuous agents, and an agent marketplace | 185k | Python/TypeScript | Significant Gravitas |
+| 3 | [n8n](https://github.com/n8n-io/n8n) | Workflow automation, native AI features, 400+ integrations, visual/code hybrid | 175k | TypeScript | n8n.io |
+| 4 | [LangChain / LangGraph](https://github.com/langchain-ai/langchain) | Full ecosystem: LangChain for chains/model+tool abstractions, [LangGraph](https://github.com/langchain-ai/langgraph) (36.7k) for graph-native orchestration with explicit state, branching, loops, checkpoints, human-in-the-loop, durable execution, streaming; [LangGraphJS](https://github.com/langchain-ai/langgraphjs) (3.1k) mirrors it for JS runtimes | 141k | Python + TS/JS | LangChain AI |
+| 5 | [Dify](https://github.com/langgenius/dify) | Visual workflow builder, RAG pipelines, agent apps, deployment tooling | 130k–147k+ | TypeScript | LangGenius |
+| 6 | [OpenHands](https://github.com/All-Hands-AI/OpenHands) | Autonomous dev agent — writes/edits code, runs shell commands, browses the web, executes in a sandbox; strong local secure-workflow story | 80k | Python | All-Hands-AI (formerly OpenDevin) |
+| 7 | [MetaGPT](https://github.com/FoundationAgents/MetaGPT) | Simulates a software company (PM/architect/dev/QA roles), SOP-driven pipelines, generates specs, docs, and code from one prompt | 69k | Python | Foundation Agents |
+| 8 | [Mem0](https://github.com/mem0ai/mem0) | Memory layer, not an orchestrator: extraction, consolidation, scoped retrieval, long-term user/agent memory; drops into any framework | 60k | Python/TypeScript | mem0ai |
+| 9 | [AutoGen](https://github.com/microsoft/autogen) | Dialogue-first multi-agent conversations, async event-driven messaging, distributed collaboration, observability, code execution | 60k | Python | Microsoft |
+| 10 | [CrewAI](https://github.com/crewAIInc/crewAI) | Role-based "crews," task delegation and collaboration, memory, checkpointing, async flows; lean (no LangChain dependency) | 55k | Python | CrewAI Inc. / João Moura |
+| 11 | [LlamaIndex](https://github.com/run-llama/llama_index) | Retrieval-centric agents, event-driven Workflows, RAG-first data connectors, memory, structured query planning | 51k | Python | LlamaIndex |
+| 12 | [smolagents](https://github.com/huggingface/smolagents) | Minimal code-first agents that write actions as Python code, sandboxed execution, tiny surface area, model-agnostic | 28k | Python | Hugging Face |
+| 13 | [OpenAI Agents SDK](https://github.com/openai/openai-agents-python) | Lightweight agent loop, handoffs between agents, guardrails, sessions, built-in tracing, sandboxed tool execution | 28k | Python/TypeScript | OpenAI |
+| 14 | [Mastra](https://github.com/mastra-ai/mastra) | TS-native, batteries-included: agents, graph-based workflows (`.then()`/`.branch()`/`.parallel()`), 40+ model routing, observational memory, human-in-the-loop, authored MCP servers, built-in evals + observability; deploys standalone or inside React/Next/Node | 26k | TypeScript | Mastra (mastra.ai) |
+| 15 | [Vercel AI SDK](https://github.com/vercel/ai) | Streaming, tool calling, structured outputs, agent loops, MCP support, provider-agnostic, first-class UI hooks; runs on edge/Workers | 25k | TypeScript | Vercel |
+| 16 | [Zep](https://github.com/getzep/zep) | Agent memory platform built on a temporal knowledge graph; long-term context, fact extraction, cross-session recall | 4.7k | Python/TypeScript | getzep |

package/package.json

@@ -0,0 +1,101 @@
+{
+  "name": "chat-agent-toolkit",
+  "version": "1.2.1",
+  "description": "Multi-provider AI agent toolkit: generate language responses, search the web, extract content, and manage memory across 10+ LLM providers.",
+  "author": "vtempest <grokthiscontact@gmail.com>",
+  "license": "AGPL-3.0",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/OpenSourceAGI/qwksearch-research-agent"
+  },
+  "bugs": {
+    "url": "https://github.com/OpenSourceAGI/qwksearch-research-agent/issues"
+  },
+  "main": "./dist/research-agent.cjs.js",
+  "types": "./dist/types.d.ts",
+  "exports": {
+    ".": {
+      "types": "./dist/types.d.ts",
+      "import": "./dist/research-agent.es.js",
+      "require": "./dist/research-agent.cjs.js"
+    },
+    "./*": {
+      "types": "./src/*.ts",
+      "react-server": "./src/*",
+      "import": "./src/*",
+      "require": "./src/*"
+    }
+  },
+  "typesVersions": {
+    "*": {
+      "*": [
+        "src/*"
+      ]
+    }
+  },
+  "files": [
+    "dist",
+    "src"
+  ],
+  "sideEffects": false,
+  "scripts": {
+    "build": "vite build",
+    "test": "vitest",
+    "test-ui": "vitest --ui --watch",
+    "make": "rm -rf dist/*; NODE_OPTIONS=--max-old-space-size=15192 vite build",
+    "ship": "npx standard-version --release-as patch && npm publish --access public"
+  },
+  "peerDependencies": {
+    "next": ">=15.0.0",
+    "vinext": ">=0.1.0 || >=1.0.0-0"
+  },
+  "peerDependenciesMeta": {
+    "next": {
+      "optional": true
+    },
+    "vinext": {
+      "optional": true
+    }
+  },
+  "devDependencies": {
+    "@vitest/ui": "^4.0.18",
+    "terser": "^5.46.0",
+    "typedoc-plugin-markdown": "^4.10.0",
+    "typescript": "^5.9.3",
+    "vite": "^8.1.3",
+    "vite-plugin-dts": "^5.0.3",
+    "vite-plugin-node-polyfills": "^0.28.0",
+    "vitest": "^4.0.18"
+  },
+  "dependencies": {
+    "@ai-sdk/anthropic": "^2.0.0",
+    "@ai-sdk/google": "^2.0.0",
+    "@ai-sdk/groq": "^2.0.0",
+    "@ai-sdk/openai": "^2.0.0",
+    "@composio/core": "^0.10.0",
+    "@mastra/core": "^0.1.0",
+    "@mastra/mcp": "^0.1.0",
+    "ai": "^5.0.0",
+    "drizzle-orm": "^0.45.1",
+    "qwksearch-api-client": "^0.0.12",
+    "zod": "^4.3.6"
+  },
+  "keywords": [
+    "ai-agent",
+    "llm",
+    "research-agent",
+    "vercel-ai-sdk",
+    "nlp",
+    "web-search",
+    "knowledge-graph",
+    "multi-provider",
+    "groq",
+    "openai",
+    "anthropic",
+    "ollama",
+    "hacktoberfest"
+  ],
+  "publishConfig": {
+    "access": "public"
+  }
+}

package/src/config/config-manager.ts

@@ -0,0 +1,319 @@
+/**
+ * @fileoverview Configuration Manager
+ *
+ * Manages model providers, MCP servers, and search configuration in memory.
+ * Handles environment variable loading, provider hashing, and config updates.
+ */
+import type { ConfigModelProvider, MCPServerConfig, Config, UIConfigSections, Model } from "./config-types";
+import { getModelProvidersUIConfigSection } from "./provider-ui-config";
+import { getEnv } from "./environment-variables";
+import { LANGUAGE_MODELS } from "./language-models-database";
+
+// Maps a provider UI key (from provider-ui-config) to the matching
+// `provider` name in the LANGUAGE_MODELS database. Names differ in a few
+// cases (e.g. the "gemini" UI key corresponds to the "Google" model list).
+const PROVIDER_KEY_TO_DB_NAME: Record<string, string> = {
+  openai: "openai",
+  anthropic: "anthropic",
+  gemini: "google",
+  groq: "groq",
+  deepseek: "deepseek",
+  nvidia: "nvidia",
+  openrouter: "openrouter",
+};
+
+/**
+ * Returns the default chat models for a provider from the LANGUAGE_MODELS
+ * database so that env-based providers expose a usable model list instead of
+ * an empty one. Returns [] when no matching provider list exists.
+ */
+const getDefaultChatModels = (providerKey: string): Model[] => {
+  const dbName = PROVIDER_KEY_TO_DB_NAME[providerKey] ?? providerKey;
+  const entry = LANGUAGE_MODELS.find(
+    (p) => p.provider.toLowerCase() === dbName.toLowerCase(),
+  );
+  if (!entry?.models) return [];
+
+  return entry.models.map((m: any) => ({
+    name: m.name,
+    key: m.id,
+  }));
+};
+
+const hashObj = (obj: { [key: string]: any }) => {
+  const str = JSON.stringify(obj, Object.keys(obj).sort());
+  let hash = 0;
+  for (let i = 0; i < str.length; i++) {
+    hash = (hash << 5) - hash + str.charCodeAt(i);
+    hash |= 0;
+  }
+  return String(Math.abs(hash).toString(36));
+};
+
+class ConfigManager {
+  configVersion = 1;
+  currentConfig: Config = {
+    version: this.configVersion,
+    setupComplete: getEnv("SETUP_COMPLETE") === "true" || false,
+    preferences: {},
+    personalization: {},
+    modelProviders: [],
+    mcpServers: [],
+    search: {
+      searxngURL: "",
+      tavilyApiKey: "",
+      sourceScrapeCount: 3,
+      sourceScrapeTimeout: 5,
+    },
+  };
+  uiConfigSections: UIConfigSections = {
+    preferences: [],
+    personalization: [],
+    modelProviders: [],
+    mcpServers: [],
+    search: [
+      {
+        name: "SearXNG URL",
+        key: "searxngURL",
+        type: "string",
+        required: false,
+        description: "The URL of your SearXNG instance",
+        placeholder: "http://localhost:4000",
+        default: "",
+        scope: "server",
+        env: "SEARXNG_API_URL",
+      },
+      {
+        name: "Tavily API Key",
+        key: "tavilyApiKey",
+        type: "string",
+        required: false,
+        description: "Your Tavily API key for enhanced search capabilities.",
+        placeholder: "tvly-...",
+        default: "",
+        scope: "server",
+        env: "TAVILY_API_KEY",
+      },
+      {
+        name: "Source pages to scrape",
+        key: "sourceScrapeCount",
+        type: "select",
+        options: [
+          { name: "Disabled (snippet only)", value: "0" },
+          { name: "1 page", value: "1" },
+          { name: "2 pages", value: "2" },
+          { name: "3 pages (default)", value: "3" },
+          { name: "5 pages", value: "5" },
+        ],
+        required: false,
+        description: "Number of top search result URLs to fully scrape.",
+        default: "3",
+        scope: "server",
+      },
+      {
+        name: "Scrape timeout (seconds)",
+        key: "sourceScrapeTimeout",
+        type: "select",
+        options: [
+          { name: "3 seconds", value: "3" },
+          { name: "5 seconds (default)", value: "5" },
+          { name: "10 seconds", value: "10" },
+          { name: "15 seconds", value: "15" },
+          { name: "20 seconds", value: "20" },
+        ],
+        required: false,
+        description: "Maximum time to wait when scraping each source URL.",
+        default: "5",
+        scope: "server",
+      },
+    ],
+  };
+
+  private initialized = false;
+
+  constructor() {
+    // Don't initialize in constructor to avoid circular dependency
+    // Initialize lazily when config is first accessed
+  }
+
+  private ensureInitialized() {
+    if (this.initialized) return;
+    this.initialize();
+    this.initialized = true;
+  }
+
+  private initialize() {
+    const providerConfigSections = getModelProvidersUIConfigSection();
+    this.uiConfigSections.modelProviders = providerConfigSections;
+
+    const newProviders: ConfigModelProvider[] = [];
+
+    // Check if OPENROUTER_API_KEY exists - prioritize OpenRouter for guests
+    const hasOpenRouter = !!getEnv("OPENROUTER_API_KEY");
+
+    providerConfigSections.forEach((provider) => {
+      // If OpenRouter is configured, ONLY load OpenRouter for guests
+      // Skip all other providers (NVIDIA, DeepSeek, etc.)
+      if (hasOpenRouter && provider.key.toLowerCase() !== "openrouter") {
+        return;
+      }
+
+      const tempConfig: Record<string, any> = {};
+      const required: string[] = [];
+
+      provider.fields.forEach((field) => {
+        tempConfig[field.key] =
+          getEnv(field.env!) || field.default || "";
+        if (field.required) required.push(field.key);
+      });
+
+      let configured = true;
+      required.forEach((r) => {
+        if (!tempConfig[r]) configured = false;
+      });
+
+      if (configured) {
+        const hash = hashObj(tempConfig);
+        const exists = this.currentConfig.modelProviders.find(
+          (p) => p.hash === hash,
+        );
+
+        if (!exists) {
+          newProviders.push({
+            id: hash,
+            name: `${provider.name}`,
+            type: provider.key,
+            chatModels: getDefaultChatModels(provider.key),
+            config: tempConfig,
+            hash: hash,
+            isEnvBased: true,
+          });
+        }
+      }
+    });
+
+    if (newProviders.length > 0) {
+      this.currentConfig.modelProviders.push(...newProviders);
+    }
+
+    // Search config from env
+    this.uiConfigSections.search.forEach((f) => {
+      if (f.env && !this.currentConfig.search[f.key]) {
+        this.currentConfig.search[f.key] =
+          getEnv(f.env) ?? f.default ?? "";
+      }
+    });
+  }
+
+  public getConfig(key: string, defaultValue?: any): any {
+    this.ensureInitialized();
+    const nested = key.split(".");
+    let obj: any = this.currentConfig;
+
+    for (let i = 0; i < nested.length; i++) {
+      const part = nested[i];
+      if (obj == null) return defaultValue;
+      obj = obj[part];
+    }
+
+    return obj === undefined ? defaultValue : obj;
+  }
+
+  public updateConfig(key: string, val: any) {
+    const parts = key.split(".");
+    if (parts.length === 0) return;
+
+    let target: any = this.currentConfig;
+    for (let i = 0; i < parts.length - 1; i++) {
+      const part = parts[i];
+      if (target[part] === null || typeof target[part] !== "object") {
+        target[part] = {};
+      }
+      target = target[part];
+    }
+
+    const finalKey = parts[parts.length - 1];
+    target[finalKey] = val;
+  }
+
+  public addModelProvider(type: string, name: string, config: any) {
+    this.ensureInitialized();
+    const hash = hashObj(config);
+
+    const newModelProvider: ConfigModelProvider = {
+      id: hash,
+      name,
+      type,
+      config,
+      chatModels: [],
+      hash: hash,
+    };
+
+    this.currentConfig.modelProviders.push(newModelProvider);
+    return newModelProvider;
+  }
+
+  public removeModelProvider(id: string) {
+    this.currentConfig.modelProviders =
+      this.currentConfig.modelProviders.filter((p) => p.id !== id);
+  }
+
+  public async updateModelProvider(id: string, name: string, config: any) {
+    const provider = this.currentConfig.modelProviders.find(
+      (p) => p.id === id,
+    );
+    if (!provider) throw new Error("Provider not found");
+
+    provider.name = name;
+    provider.config = config;
+    return provider;
+  }
+
+  public addProviderModel(providerId: string, type: "chat", model: any) {
+    const provider = this.currentConfig.modelProviders.find(
+      (p) => p.id === providerId,
+    );
+    if (!provider) throw new Error("Invalid provider id");
+
+    delete model.type;
+    provider.chatModels.push(model);
+    return model;
+  }
+
+  public removeProviderModel(
+    providerId: string,
+    type: "chat",
+    modelKey: string,
+  ) {
+    const provider = this.currentConfig.modelProviders.find(
+      (p) => p.id === providerId,
+    );
+    if (!provider) throw new Error("Invalid provider id");
+
+    provider.chatModels = provider.chatModels.filter((m) => m.key !== modelKey);
+  }
+
+  public isSetupComplete() {
+    return this.currentConfig.setupComplete;
+  }
+
+  public markSetupComplete() {
+    if (!this.currentConfig.setupComplete) {
+      this.currentConfig.setupComplete = true;
+    }
+  }
+
+  public getUIConfigSections(): UIConfigSections {
+    this.ensureInitialized();
+    return this.uiConfigSections;
+  }
+
+  public getCurrentConfig(): Config {
+    this.ensureInitialized();
+    return JSON.parse(JSON.stringify(this.currentConfig));
+  }
+}
+
+const configManager = new ConfigManager();
+
+export default configManager;