@graphorin/provider 0.6.0 → 0.6.1

package/CHANGELOG.md

@@ -1,5 +1,15 @@
 # @graphorin/provider
 
+## 0.6.1
+
+### Patch Changes
+
+- [#142](https://github.com/o-stepper/graphorin/pull/142) [`436d6ca`](https://github.com/o-stepper/graphorin/commit/436d6ca5ebbd16df094e915682d3915c279a8430) Thanks [@o-stepper](https://github.com/o-stepper)! - Version constants and version-bearing strings now derive from each package's manifest at build time (`VERSION = pkg.version`; writer ids, client/server info, OTLP framework attributes, build-info metrics interpolate it). No behavioral change at the current version: the rendered strings are byte-identical. A release bump no longer edits source; the new `check-version-consistency` gate fails any reintroduced hardcoded framework version.
+
+- Updated dependencies [[`826ee59`](https://github.com/o-stepper/graphorin/commit/826ee5933ad38693b0dd2f20a110abfecba7d23d), [`436d6ca`](https://github.com/o-stepper/graphorin/commit/436d6ca5ebbd16df094e915682d3915c279a8430)]:
+  - @graphorin/observability@0.6.1
+  - @graphorin/core@0.6.1
+
 ## 0.6.0
 
 ### Minor Changes

package/README.md

@@ -11,11 +11,16 @@ The package owns four moving parts:
    AI SDK), `ollamaAdapter` (direct Ollama HTTP), `llamaCppServerAdapter`
    (the upstream `llama-server` binary from llama.cpp), and
    `openAICompatibleAdapter` (LMStudio / LocalAI / vLLM / Together-style
-   self-host endpoints).
+   self-host endpoints). Prompt caching: when a request carries
+   `cachePolicy: { breakpoints: 'auto' }`, the vercel adapter anchors
+   Anthropic `cache_control` breakpoints on the stable conversation
+   prefix, and cache read / write token legs flow back through
+   `Usage.cachedReadTokens` / `cacheWriteTokens`.
 3. **Middleware** - `composeProviderMiddleware([...])` enforces a
    canonical order at startup and throws `MiddlewareOrderingError` on
    violation. Built-ins: `withTracing`, `withRetry`, `withRateLimit`,
-   `withCostLimit`, `withCostTracking`, `withFallback`, and
+   `withCostLimit`, `withCostTracking` (cache-aware: bills cache reads
+   and writes at their own rates), `withFallback`, and
    `withRedaction` (mandatory in production).
 4. **Token counting** - pluggable `TokenCounter` dispatcher. Default
    `JsTiktokenCounter` for OpenAI-compatible models; per-vendor native
@@ -101,5 +106,5 @@ recommendations.
 
 ## Project metadata
 
-- **Project Graphorin** · v0.6.0 · MIT License · © 2026 Oleksiy Stepurenko
+- **Project Graphorin** · v0.6.1 · MIT License · © 2026 Oleksiy Stepurenko
 - Repository: <https://github.com/o-stepper/graphorin>

package/dist/index.d.ts

@@ -63,8 +63,7 @@ import { LocalProviderTrust, OllamaTrust } from "./trust/index.js";
  *
  * @packageDocumentation
  */
-/** Canonical version constant. Mirrors the `package.json` version. */
-declare const VERSION = "0.6.0";
+declare const VERSION: string;
 //#endregion
 export { AnthropicAPICounter, AnthropicAPICounterOptions, ApplyReasoningPolicyInput, BedrockAPICounter, BedrockAPICounterOptions, CANONICAL_MIDDLEWARE_ORDER, CLASSIFIER_RULES, ClassifierRule, CostAccumulator, CostBudgetExceededError, CostTrackingTotals, CreateDefaultCounterOptions, CreateProviderOptions, GoogleAPICounter, GoogleAPICounterOptions, GraphorinProviderError, HeuristicCounter, HeuristicCounterOptions, InferReasoningContractInput, InvalidProviderError, JsTiktokenCounter, JsTiktokenCounterOptions, LlamaCppServerAdapterOptions, LocalProviderClassification, LocalProviderInsecureTransportError, LocalProviderTrust, MIDDLEWARE_KIND, MiddlewareOrderingError, MissingProductionMiddlewareError, OllamaAdapterOptions, OllamaInsecureTransportError, OllamaTrust, OpenAICompatibleAdapterOptions, PERMANENT_LOOPBACK_CLASSIFICATION, ProductionStartupHookOptions, PromptRedactionError, PromptRedactionPolicy, PromptRedactionScanScope, PromptRedactionViolation, ProviderHttpError, ProviderStreamParseError, REASONING_CONTRACT_RULES, REASONING_RETENTION_DEFAULTS, RateLimitExceededError, ReasoningContractRule, ResolveReasoningRetentionInput, SENSITIVITY_DEFAULTS_PER_TRUST, SerializedMessage, VERSION, VercelAdapterOptions, WithCostLimitOptions, WithCostTrackingOptions, WithFallbackOptions, WithRateLimitOptions, WithRetryOptions, WithTracingOptions, __resetGlobalTokenCounter, __resetTiktokenCache, applyReasoningPolicy, assertProductionMiddleware, classifyHttpStatus, classifyLocalProvider, classifyModelTier, composeProviderMiddleware, createCostAccumulator, createDefaultCounter, createProvider, defineProviderMiddleware, foldToolExamples, getGlobalTokenCounter, getMiddlewareKind, inferReasoningContract, llamaCppServerAdapter, ollamaAdapter, openAICompatibleAdapter, providerHasMiddleware, resolveReasoningRetention, serialiseMessageForCount, serializedToString, setGlobalTokenCounter, vercelAdapter, withCostLimit, withCostTracking, withFallback, withRateLimit, withRedaction, withRetry, withTracing };
 //# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","names":[],"sources":["../src/index.ts"],"sourcesContent":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;cA6Ba,OAAA"}
+{"version":3,"file":"index.d.ts","names":[],"sources":["../src/index.ts"],"sourcesContent":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;cA+Ba"}

package/dist/index.js

@@ -1,3 +1,4 @@
+import { version } from "./package.js";
 import { CostBudgetExceededError, GraphorinProviderError, InvalidProviderError, LocalProviderInsecureTransportError, MiddlewareOrderingError, MissingProductionMiddlewareError, OllamaInsecureTransportError, PromptRedactionError, ProviderHttpError, ProviderStreamParseError, RateLimitExceededError, classifyHttpStatus } from "./errors/errors.js";
 import { applyReasoningPolicy } from "./reasoning/apply-policy.js";
 import { REASONING_RETENTION_DEFAULTS, resolveReasoningRetention } from "./reasoning/retention.js";
@@ -62,8 +63,8 @@ import "./trust/index.js";
 *
 * @packageDocumentation
 */
-/** Canonical version constant. Mirrors the `package.json` version. */
-const VERSION = "0.6.0";
+/** Canonical version constant, derived from `package.json` at build time. */
+const VERSION = version;
 
 //#endregion
 export { AnthropicAPICounter, BedrockAPICounter, CANONICAL_MIDDLEWARE_ORDER, CLASSIFIER_RULES, CostBudgetExceededError, GoogleAPICounter, GraphorinProviderError, HeuristicCounter, InvalidProviderError, JsTiktokenCounter, LocalProviderInsecureTransportError, MIDDLEWARE_KIND, MiddlewareOrderingError, MissingProductionMiddlewareError, OllamaInsecureTransportError, PERMANENT_LOOPBACK_CLASSIFICATION, PromptRedactionError, ProviderHttpError, ProviderStreamParseError, REASONING_CONTRACT_RULES, REASONING_RETENTION_DEFAULTS, RateLimitExceededError, SENSITIVITY_DEFAULTS_PER_TRUST, VERSION, __resetGlobalTokenCounter, __resetTiktokenCache, applyReasoningPolicy, assertProductionMiddleware, classifyHttpStatus, classifyLocalProvider, classifyModelTier, composeProviderMiddleware, createCostAccumulator, createDefaultCounter, createProvider, defineProviderMiddleware, foldToolExamples, getGlobalTokenCounter, getMiddlewareKind, inferReasoningContract, llamaCppServerAdapter, ollamaAdapter, openAICompatibleAdapter, providerHasMiddleware, resolveReasoningRetention, serialiseMessageForCount, serializedToString, setGlobalTokenCounter, vercelAdapter, withCostLimit, withCostTracking, withFallback, withRateLimit, withRedaction, withRetry, withTracing };

package/dist/index.js.map

@@ -1 +1 @@
-{"version":3,"file":"index.js","names":[],"sources":["../src/index.ts"],"sourcesContent":["/**\n * @graphorin/provider - vendor-neutral LLM provider layer for the\n * Graphorin framework.\n *\n * The package owns:\n *\n * - The `createProvider(...)` factory for wrapping adapter modules in\n *   a stable `Provider` shape.\n * - Four adapters out of the box: the Vercel AI SDK adapter for the\n *   default cloud path, plus three local-LLM adapters (Ollama HTTP,\n *   llama-server HTTP, and a generic OpenAI-compatible adapter for\n *   LMStudio / LocalAI / vLLM / Together-style endpoints). All three\n *   `baseUrl`-driven adapters share the same {@link LocalProviderTrust}\n *   classifier.\n * - The canonical-order middleware composer\n *   ({@link composeProviderMiddleware}) and the seven built-in\n *   middlewares (`withTracing`, `withRetry`, `withRateLimit`,\n *   `withCostLimit`, `withCostTracking`, `withFallback`,\n *   `withRedaction`).\n * - The pluggable `TokenCounter` dispatcher (`createDefaultCounter`)\n *   plus the per-vendor strategies and a heuristic fallback.\n * - The per-provider model-tier auto-classifier\n *   ({@link classifyModelTier}) that maps a `Provider`'s `modelId` to\n *   one of `'fast' | 'balanced' | 'smart'`.\n *\n * @packageDocumentation\n */\n\n/** Canonical version constant. Mirrors the `package.json` version. */\nexport const VERSION = '0.6.0';\n\nexport * from './adapters/index.js';\nexport * from './counters/index.js';\nexport * from './errors/index.js';\nexport * from './middleware/index.js';\nexport * from './model-tier/index.js';\nexport * from './provider.js';\nexport * from './reasoning/index.js';\nexport { foldToolExamples } from './tool-examples.js';\nexport * from './trust/index.js';\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AA6BA,MAAa,UAAU"}
+{"version":3,"file":"index.js","names":["VERSION: string","pkg.version"],"sources":["../src/index.ts"],"sourcesContent":["/**\n * @graphorin/provider - vendor-neutral LLM provider layer for the\n * Graphorin framework.\n *\n * The package owns:\n *\n * - The `createProvider(...)` factory for wrapping adapter modules in\n *   a stable `Provider` shape.\n * - Four adapters out of the box: the Vercel AI SDK adapter for the\n *   default cloud path, plus three local-LLM adapters (Ollama HTTP,\n *   llama-server HTTP, and a generic OpenAI-compatible adapter for\n *   LMStudio / LocalAI / vLLM / Together-style endpoints). All three\n *   `baseUrl`-driven adapters share the same {@link LocalProviderTrust}\n *   classifier.\n * - The canonical-order middleware composer\n *   ({@link composeProviderMiddleware}) and the seven built-in\n *   middlewares (`withTracing`, `withRetry`, `withRateLimit`,\n *   `withCostLimit`, `withCostTracking`, `withFallback`,\n *   `withRedaction`).\n * - The pluggable `TokenCounter` dispatcher (`createDefaultCounter`)\n *   plus the per-vendor strategies and a heuristic fallback.\n * - The per-provider model-tier auto-classifier\n *   ({@link classifyModelTier}) that maps a `Provider`'s `modelId` to\n *   one of `'fast' | 'balanced' | 'smart'`.\n *\n * @packageDocumentation\n */\n\n/** Canonical version constant, derived from `package.json` at build time. */\nimport pkg from '../package.json' with { type: 'json' };\n\nexport const VERSION: string = pkg.version;\n\nexport * from './adapters/index.js';\nexport * from './counters/index.js';\nexport * from './errors/index.js';\nexport * from './middleware/index.js';\nexport * from './model-tier/index.js';\nexport * from './provider.js';\nexport * from './reasoning/index.js';\nexport { foldToolExamples } from './tool-examples.js';\nexport * from './trust/index.js';\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AA+BA,MAAaA,UAAkBC"}

package/dist/package.js

@@ -0,0 +1,6 @@
+//#region package.json
+var version = "0.6.1";
+
+//#endregion
+export { version };
+//# sourceMappingURL=package.js.map

package/dist/package.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"package.js","names":[],"sources":["../package.json"],"sourcesContent":["{\n  \"name\": \"@graphorin/provider\",\n  \"version\": \"0.6.1\",\n  \"description\": \"Vendor-neutral LLM provider layer for the Graphorin framework. Ships the canonical Provider interface, the Vercel AI SDK adapter (default cloud path), three local-LLM adapters (direct Ollama HTTP, llama-server HTTP, and a generic OpenAI-compatible adapter for LMStudio / LocalAI / vLLM / Together-style endpoints), the shared LocalProviderTrust classifier (loopback / private / public-tls / public-cleartext), the canonical-order middleware composer (withTracing / withRetry / withRateLimit / withCostLimit / withCostTracking / withFallback / withRedaction), the pluggable TokenCounter dispatcher (js-tiktoken default plus per-vendor native counters and a heuristic fallback), the per-provider model-tier auto-classifier (fast / balanced / smart), and the production startup hook that fails fast when withRedaction is missing.\",\n  \"license\": \"MIT\",\n  \"author\": \"Oleksiy Stepurenko\",\n  \"homepage\": \"https://github.com/o-stepper/graphorin/tree/main/packages/provider\",\n  \"repository\": {\n    \"type\": \"git\",\n    \"url\": \"git+https://github.com/o-stepper/graphorin.git\",\n    \"directory\": \"packages/provider\"\n  },\n  \"bugs\": {\n    \"url\": \"https://github.com/o-stepper/graphorin/issues\"\n  },\n  \"keywords\": [\n    \"graphorin\",\n    \"ai\",\n    \"agents\",\n    \"framework\",\n    \"provider\",\n    \"llm\",\n    \"vercel-ai-sdk\",\n    \"ollama\",\n    \"llama-cpp\",\n    \"openai-compatible\",\n    \"middleware\",\n    \"retry\",\n    \"fallback\",\n    \"rate-limit\",\n    \"cost-limit\",\n    \"redaction\",\n    \"token-counter\",\n    \"tiktoken\",\n    \"local-first\"\n  ],\n  \"type\": \"module\",\n  \"engines\": {\n    \"node\": \">=22.0.0\"\n  },\n  \"main\": \"./dist/index.js\",\n  \"module\": \"./dist/index.js\",\n  \"types\": \"./dist/index.d.ts\",\n  \"exports\": {\n    \".\": {\n      \"types\": \"./dist/index.d.ts\",\n      \"import\": \"./dist/index.js\"\n    },\n    \"./adapters/vercel\": {\n      \"types\": \"./dist/adapters/vercel.d.ts\",\n      \"import\": \"./dist/adapters/vercel.js\"\n    },\n    \"./adapters/ollama\": {\n      \"types\": \"./dist/adapters/ollama.d.ts\",\n      \"import\": \"./dist/adapters/ollama.js\"\n    },\n    \"./adapters/llamacpp-server\": {\n      \"types\": \"./dist/adapters/llamacpp-server.d.ts\",\n      \"import\": \"./dist/adapters/llamacpp-server.js\"\n    },\n    \"./adapters/openai-compatible\": {\n      \"types\": \"./dist/adapters/openai-compatible.d.ts\",\n      \"import\": \"./dist/adapters/openai-compatible.js\"\n    },\n    \"./middleware\": {\n      \"types\": \"./dist/middleware/index.d.ts\",\n      \"import\": \"./dist/middleware/index.js\"\n    },\n    \"./counters\": {\n      \"types\": \"./dist/counters/index.d.ts\",\n      \"import\": \"./dist/counters/index.js\"\n    },\n    \"./trust\": {\n      \"types\": \"./dist/trust/index.d.ts\",\n      \"import\": \"./dist/trust/index.js\"\n    },\n    \"./model-tier\": {\n      \"types\": \"./dist/model-tier/index.d.ts\",\n      \"import\": \"./dist/model-tier/index.js\"\n    },\n    \"./reasoning\": {\n      \"types\": \"./dist/reasoning/index.d.ts\",\n      \"import\": \"./dist/reasoning/index.js\"\n    },\n    \"./errors\": {\n      \"types\": \"./dist/errors/index.d.ts\",\n      \"import\": \"./dist/errors/index.js\"\n    },\n    \"./package.json\": \"./package.json\"\n  },\n  \"files\": [\n    \"dist\",\n    \"README.md\",\n    \"CHANGELOG.md\",\n    \"LICENSE\"\n  ],\n  \"scripts\": {\n    \"build\": \"tsdown\",\n    \"typecheck\": \"tsc --noEmit && tsc -p tsconfig.tests.json\",\n    \"test\": \"vitest run\",\n    \"lint\": \"biome check .\",\n    \"clean\": \"rimraf dist .turbo *.tsbuildinfo\"\n  },\n  \"dependencies\": {\n    \"@graphorin/core\": \"workspace:*\",\n    \"@graphorin/observability\": \"workspace:*\"\n  },\n  \"peerDependencies\": {\n    \"ai\": \"^7.0.0-beta.76\",\n    \"js-tiktoken\": \"^1.0.0\"\n  },\n  \"peerDependenciesMeta\": {\n    \"ai\": {\n      \"optional\": true\n    },\n    \"js-tiktoken\": {\n      \"optional\": true\n    }\n  },\n  \"publishConfig\": {\n    \"access\": \"public\",\n    \"provenance\": true\n  },\n  \"devDependencies\": {\n    \"@graphorin/pricing\": \"workspace:*\",\n    \"ai\": \"^7.0.15\",\n    \"fast-check\": \"^3.23.0\",\n    \"js-tiktoken\": \"^1.0.21\"\n  }\n}\n"],"mappings":";cAEa"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@graphorin/provider",
-  "version": "0.6.0",
+  "version": "0.6.1",
   "description": "Vendor-neutral LLM provider layer for the Graphorin framework. Ships the canonical Provider interface, the Vercel AI SDK adapter (default cloud path), three local-LLM adapters (direct Ollama HTTP, llama-server HTTP, and a generic OpenAI-compatible adapter for LMStudio / LocalAI / vLLM / Together-style endpoints), the shared LocalProviderTrust classifier (loopback / private / public-tls / public-cleartext), the canonical-order middleware composer (withTracing / withRetry / withRateLimit / withCostLimit / withCostTracking / withFallback / withRedaction), the pluggable TokenCounter dispatcher (js-tiktoken default plus per-vendor native counters and a heuristic fallback), the per-provider model-tier auto-classifier (fast / balanced / smart), and the production startup hook that fails fast when withRedaction is missing.",
   "license": "MIT",
   "author": "Oleksiy Stepurenko",
@@ -95,8 +95,8 @@
     "LICENSE"
   ],
   "dependencies": {
-    "@graphorin/core": "0.6.0",
-    "@graphorin/observability": "0.6.0"
+    "@graphorin/core": "0.6.1",
+    "@graphorin/observability": "0.6.1"
   },
   "peerDependencies": {
     "ai": "^7.0.0-beta.76",
@@ -118,7 +118,7 @@
     "ai": "^7.0.15",
     "fast-check": "^3.23.0",
     "js-tiktoken": "^1.0.21",
-    "@graphorin/pricing": "0.6.0"
+    "@graphorin/pricing": "0.6.1"
   },
   "scripts": {
     "build": "tsdown",