phi-code-ai 0.56.3 → 0.74.0

package/dist/models.js

@@ -27,21 +27,37 @@ export function calculateCost(model, usage) {
     usage.cost.total = usage.cost.input + usage.cost.output + usage.cost.cacheRead + usage.cost.cacheWrite;
     return usage.cost;
 }
-/**
- * Check if a model supports xhigh thinking level.
- *
- * Supported today:
- * - GPT-5.2 / GPT-5.3 / GPT-5.4 model families
- * - Anthropic Messages API Opus 4.6 models (xhigh maps to adaptive effort "max")
- */
-export function supportsXhigh(model) {
-    if (model.id.includes("gpt-5.2") || model.id.includes("gpt-5.3") || model.id.includes("gpt-5.4")) {
+const EXTENDED_THINKING_LEVELS = ["off", "minimal", "low", "medium", "high", "xhigh"];
+export function getSupportedThinkingLevels(model) {
+    if (!model.reasoning)
+        return ["off"];
+    return EXTENDED_THINKING_LEVELS.filter((level) => {
+        const mapped = model.thinkingLevelMap?.[level];
+        if (mapped === null)
+            return false;
+        if (level === "xhigh")
+            return mapped !== undefined;
         return true;
+    });
+}
+export function clampThinkingLevel(model, level) {
+    const availableLevels = getSupportedThinkingLevels(model);
+    if (availableLevels.includes(level))
+        return level;
+    const requestedIndex = EXTENDED_THINKING_LEVELS.indexOf(level);
+    if (requestedIndex === -1)
+        return availableLevels[0] ?? "off";
+    for (let i = requestedIndex; i < EXTENDED_THINKING_LEVELS.length; i++) {
+        const candidate = EXTENDED_THINKING_LEVELS[i];
+        if (availableLevels.includes(candidate))
+            return candidate;
     }
-    if (model.api === "anthropic-messages") {
-        return model.id.includes("opus-4-6") || model.id.includes("opus-4.6");
+    for (let i = requestedIndex - 1; i >= 0; i--) {
+        const candidate = EXTENDED_THINKING_LEVELS[i];
+        if (availableLevels.includes(candidate))
+            return candidate;
     }
-    return false;
+    return availableLevels[0] ?? "off";
 }
 /**
  * Check if two models are equal by comparing both their id and provider.

package/dist/models.js.map

@@ -1 +1 @@
-{"version":3,"file":"models.js","sourceRoot":"","sources":["../src/models.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,EAAE,MAAM,uBAAuB,CAAC;AAG/C,MAAM,aAAa,GAAyC,IAAI,GAAG,EAAE,CAAC;AAEtE,iDAAiD;AACjD,KAAK,MAAM,CAAC,QAAQ,EAAE,MAAM,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;IACzD,MAAM,cAAc,GAAG,IAAI,GAAG,EAAsB,CAAC;IACrD,KAAK,MAAM,CAAC,EAAE,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;QAClD,cAAc,CAAC,GAAG,CAAC,EAAE,EAAE,KAAmB,CAAC,CAAC;IAC7C,CAAC;IACD,aAAa,CAAC,GAAG,CAAC,QAAQ,EAAE,cAAc,CAAC,CAAC;AAC7C,CAAC;AAOD,MAAM,UAAU,QAAQ,CACvB,QAAmB,EACnB,OAAiB;IAEjB,MAAM,cAAc,GAAG,aAAa,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;IACnD,OAAO,cAAc,EAAE,GAAG,CAAC,OAAiB,CAAyC,CAAC;AACvF,CAAC;AAED,MAAM,UAAU,YAAY;IAC3B,OAAO,KAAK,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,EAAE,CAAoB,CAAC;AAC5D,CAAC;AAED,MAAM,UAAU,SAAS,CACxB,QAAmB;IAEnB,MAAM,MAAM,GAAG,aAAa,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;IAC3C,OAAO,MAAM,CAAC,CAAC,CAAE,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAoE,CAAC,CAAC,CAAC,EAAE,CAAC;AACtH,CAAC;AAED,MAAM,UAAU,aAAa,CAAmB,KAAkB,EAAE,KAAY;IAC/E,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,OAAO,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC;IAC9D,KAAK,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,GAAG,KAAK,CAAC,MAAM,CAAC;IACjE,KAAK,CAAC,IAAI,CAAC,SAAS,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,SAAS,GAAG,OAAO,CAAC,GAAG,KAAK,CAAC,SAAS,CAAC;IAC1E,KAAK,CAAC,IAAI,CAAC,UAAU,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,UAAU,GAAG,OAAO,CAAC,GAAG,KAAK,CAAC,UAAU,CAAC;IAC7E,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC,IAAI,CAAC,SAAS,GAAG,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC;IACvG,OAAO,KAAK,CAAC,IAAI,CAAC;AACnB,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,aAAa,CAAmB,KAAkB;IACjE,IAAI,KAAK,CAAC,EAAE,CAAC,QAAQ,CAAC,SAAS,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC,QAAQ,CAAC,SAAS,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,CAAC;QAClG,OAAO,IAAI,CAAC;IACb,CAAC;IAED,IAAI,KAAK,CAAC,GAAG,KAAK,oBAAoB,EAAE,CAAC;QACxC,OAAO,KAAK,CAAC,EAAE,CAAC,QAAQ,CAAC,UAAU,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC;IACvE,CAAC;IAED,OAAO,KAAK,CAAC;AACd,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,cAAc,CAC7B,CAAiC,EACjC,CAAiC;IAEjC,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC;QAAE,OAAO,KAAK,CAAC;IAC3B,OAAO,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,QAAQ,KAAK,CAAC,CAAC,QAAQ,CAAC;AACnD,CAAC","sourcesContent":["import { MODELS } from \"./models.generated.js\";\nimport type { Api, KnownProvider, Model, Usage } from \"./types.js\";\n\nconst modelRegistry: Map<string, Map<string, Model<Api>>> = new Map();\n\n// Initialize registry from MODELS on module load\nfor (const [provider, models] of Object.entries(MODELS)) {\n\tconst providerModels = new Map<string, Model<Api>>();\n\tfor (const [id, model] of Object.entries(models)) {\n\t\tproviderModels.set(id, model as Model<Api>);\n\t}\n\tmodelRegistry.set(provider, providerModels);\n}\n\ntype ModelApi<\n\tTProvider extends KnownProvider,\n\tTModelId extends keyof (typeof MODELS)[TProvider],\n> = (typeof MODELS)[TProvider][TModelId] extends { api: infer TApi } ? (TApi extends Api ? TApi : never) : never;\n\nexport function getModel<TProvider extends KnownProvider, TModelId extends keyof (typeof MODELS)[TProvider]>(\n\tprovider: TProvider,\n\tmodelId: TModelId,\n): Model<ModelApi<TProvider, TModelId>> {\n\tconst providerModels = modelRegistry.get(provider);\n\treturn providerModels?.get(modelId as string) as Model<ModelApi<TProvider, TModelId>>;\n}\n\nexport function getProviders(): KnownProvider[] {\n\treturn Array.from(modelRegistry.keys()) as KnownProvider[];\n}\n\nexport function getModels<TProvider extends KnownProvider>(\n\tprovider: TProvider,\n): Model<ModelApi<TProvider, keyof (typeof MODELS)[TProvider]>>[] {\n\tconst models = modelRegistry.get(provider);\n\treturn models ? (Array.from(models.values()) as Model<ModelApi<TProvider, keyof (typeof MODELS)[TProvider]>>[]) : [];\n}\n\nexport function calculateCost<TApi extends Api>(model: Model<TApi>, usage: Usage): Usage[\"cost\"] {\n\tusage.cost.input = (model.cost.input / 1000000) * usage.input;\n\tusage.cost.output = (model.cost.output / 1000000) * usage.output;\n\tusage.cost.cacheRead = (model.cost.cacheRead / 1000000) * usage.cacheRead;\n\tusage.cost.cacheWrite = (model.cost.cacheWrite / 1000000) * usage.cacheWrite;\n\tusage.cost.total = usage.cost.input + usage.cost.output + usage.cost.cacheRead + usage.cost.cacheWrite;\n\treturn usage.cost;\n}\n\n/**\n * Check if a model supports xhigh thinking level.\n *\n * Supported today:\n * - GPT-5.2 / GPT-5.3 / GPT-5.4 model families\n * - Anthropic Messages API Opus 4.6 models (xhigh maps to adaptive effort \"max\")\n */\nexport function supportsXhigh<TApi extends Api>(model: Model<TApi>): boolean {\n\tif (model.id.includes(\"gpt-5.2\") || model.id.includes(\"gpt-5.3\") || model.id.includes(\"gpt-5.4\")) {\n\t\treturn true;\n\t}\n\n\tif (model.api === \"anthropic-messages\") {\n\t\treturn model.id.includes(\"opus-4-6\") || model.id.includes(\"opus-4.6\");\n\t}\n\n\treturn false;\n}\n\n/**\n * Check if two models are equal by comparing both their id and provider.\n * Returns false if either model is null or undefined.\n */\nexport function modelsAreEqual<TApi extends Api>(\n\ta: Model<TApi> | null | undefined,\n\tb: Model<TApi> | null | undefined,\n): boolean {\n\tif (!a || !b) return false;\n\treturn a.id === b.id && a.provider === b.provider;\n}\n"]}
+{"version":3,"file":"models.js","sourceRoot":"","sources":["../src/models.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,EAAE,MAAM,uBAAuB,CAAC;AAG/C,MAAM,aAAa,GAAyC,IAAI,GAAG,EAAE,CAAC;AAEtE,iDAAiD;AACjD,KAAK,MAAM,CAAC,QAAQ,EAAE,MAAM,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;IACzD,MAAM,cAAc,GAAG,IAAI,GAAG,EAAsB,CAAC;IACrD,KAAK,MAAM,CAAC,EAAE,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;QAClD,cAAc,CAAC,GAAG,CAAC,EAAE,EAAE,KAAmB,CAAC,CAAC;IAC7C,CAAC;IACD,aAAa,CAAC,GAAG,CAAC,QAAQ,EAAE,cAAc,CAAC,CAAC;AAC7C,CAAC;AAOD,MAAM,UAAU,QAAQ,CACvB,QAAmB,EACnB,OAAiB,EACsB;IACvC,MAAM,cAAc,GAAG,aAAa,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;IACnD,OAAO,cAAc,EAAE,GAAG,CAAC,OAAiB,CAAyC,CAAC;AAAA,CACtF;AAED,MAAM,UAAU,YAAY,GAAoB;IAC/C,OAAO,KAAK,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,EAAE,CAAoB,CAAC;AAAA,CAC3D;AAED,MAAM,UAAU,SAAS,CACxB,QAAmB,EAC8C;IACjE,MAAM,MAAM,GAAG,aAAa,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;IAC3C,OAAO,MAAM,CAAC,CAAC,CAAE,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAoE,CAAC,CAAC,CAAC,EAAE,CAAC;AAAA,CACrH;AAED,MAAM,UAAU,aAAa,CAAmB,KAAkB,EAAE,KAAY,EAAiB;IAChG,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,OAAO,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC;IAC9D,KAAK,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,GAAG,KAAK,CAAC,MAAM,CAAC;IACjE,KAAK,CAAC,IAAI,CAAC,SAAS,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,SAAS,GAAG,OAAO,CAAC,GAAG,KAAK,CAAC,SAAS,CAAC;IAC1E,KAAK,CAAC,IAAI,CAAC,UAAU,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,UAAU,GAAG,OAAO,CAAC,GAAG,KAAK,CAAC,UAAU,CAAC;IAC7E,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC,IAAI,CAAC,SAAS,GAAG,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC;IACvG,OAAO,KAAK,CAAC,IAAI,CAAC;AAAA,CAClB;AAED,MAAM,wBAAwB,GAAyB,CAAC,KAAK,EAAE,SAAS,EAAE,KAAK,EAAE,QAAQ,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC;AAE5G,MAAM,UAAU,0BAA0B,CAAmB,KAAkB,EAAwB;IACtG,IAAI,CAAC,KAAK,CAAC,SAAS;QAAE,OAAO,CAAC,KAAK,CAAC,CAAC;IAErC,OAAO,wBAAwB,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC;QACjD,MAAM,MAAM,GAAG,KAAK,CAAC,gBAAgB,EAAE,CAAC,KAAK,CAAC,CAAC;QAC/C,IAAI,MAAM,KAAK,IAAI;YAAE,OAAO,KAAK,CAAC;QAClC,IAAI,KAAK,KAAK,OAAO;YAAE,OAAO,MAAM,KAAK,SAAS,CAAC;QACnD,OAAO,IAAI,CAAC;IAAA,CACZ,CAAC,CAAC;AAAA,CACH;AAED,MAAM,UAAU,kBAAkB,CACjC,KAAkB,EAClB,KAAyB,EACJ;IACrB,MAAM,eAAe,GAAG,0BAA0B,CAAC,KAAK,CAAC,CAAC;IAC1D,IAAI,eAAe,CAAC,QAAQ,CAAC,KAAK,CAAC;QAAE,OAAO,KAAK,CAAC;IAElD,MAAM,cAAc,GAAG,wBAAwB,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;IAC/D,IAAI,cAAc,KAAK,CAAC,CAAC;QAAE,OAAO,eAAe,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC;IAE9D,KAAK,IAAI,CAAC,GAAG,cAAc,EAAE,CAAC,GAAG,wBAAwB,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACvE,MAAM,SAAS,GAAG,wBAAwB,CAAC,CAAC,CAAC,CAAC;QAC9C,IAAI,eAAe,CAAC,QAAQ,CAAC,SAAS,CAAC;YAAE,OAAO,SAAS,CAAC;IAC3D,CAAC;IACD,KAAK,IAAI,CAAC,GAAG,cAAc,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC9C,MAAM,SAAS,GAAG,wBAAwB,CAAC,CAAC,CAAC,CAAC;QAC9C,IAAI,eAAe,CAAC,QAAQ,CAAC,SAAS,CAAC;YAAE,OAAO,SAAS,CAAC;IAC3D,CAAC;IACD,OAAO,eAAe,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC;AAAA,CACnC;AAED;;;GAGG;AACH,MAAM,UAAU,cAAc,CAC7B,CAAiC,EACjC,CAAiC,EACvB;IACV,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC;QAAE,OAAO,KAAK,CAAC;IAC3B,OAAO,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,QAAQ,KAAK,CAAC,CAAC,QAAQ,CAAC;AAAA,CAClD","sourcesContent":["import { MODELS } from \"./models.generated.js\";\nimport type { Api, KnownProvider, Model, ModelThinkingLevel, Usage } from \"./types.js\";\n\nconst modelRegistry: Map<string, Map<string, Model<Api>>> = new Map();\n\n// Initialize registry from MODELS on module load\nfor (const [provider, models] of Object.entries(MODELS)) {\n\tconst providerModels = new Map<string, Model<Api>>();\n\tfor (const [id, model] of Object.entries(models)) {\n\t\tproviderModels.set(id, model as Model<Api>);\n\t}\n\tmodelRegistry.set(provider, providerModels);\n}\n\ntype ModelApi<\n\tTProvider extends KnownProvider,\n\tTModelId extends keyof (typeof MODELS)[TProvider],\n> = (typeof MODELS)[TProvider][TModelId] extends { api: infer TApi } ? (TApi extends Api ? TApi : never) : never;\n\nexport function getModel<TProvider extends KnownProvider, TModelId extends keyof (typeof MODELS)[TProvider]>(\n\tprovider: TProvider,\n\tmodelId: TModelId,\n): Model<ModelApi<TProvider, TModelId>> {\n\tconst providerModels = modelRegistry.get(provider);\n\treturn providerModels?.get(modelId as string) as Model<ModelApi<TProvider, TModelId>>;\n}\n\nexport function getProviders(): KnownProvider[] {\n\treturn Array.from(modelRegistry.keys()) as KnownProvider[];\n}\n\nexport function getModels<TProvider extends KnownProvider>(\n\tprovider: TProvider,\n): Model<ModelApi<TProvider, keyof (typeof MODELS)[TProvider]>>[] {\n\tconst models = modelRegistry.get(provider);\n\treturn models ? (Array.from(models.values()) as Model<ModelApi<TProvider, keyof (typeof MODELS)[TProvider]>>[]) : [];\n}\n\nexport function calculateCost<TApi extends Api>(model: Model<TApi>, usage: Usage): Usage[\"cost\"] {\n\tusage.cost.input = (model.cost.input / 1000000) * usage.input;\n\tusage.cost.output = (model.cost.output / 1000000) * usage.output;\n\tusage.cost.cacheRead = (model.cost.cacheRead / 1000000) * usage.cacheRead;\n\tusage.cost.cacheWrite = (model.cost.cacheWrite / 1000000) * usage.cacheWrite;\n\tusage.cost.total = usage.cost.input + usage.cost.output + usage.cost.cacheRead + usage.cost.cacheWrite;\n\treturn usage.cost;\n}\n\nconst EXTENDED_THINKING_LEVELS: ModelThinkingLevel[] = [\"off\", \"minimal\", \"low\", \"medium\", \"high\", \"xhigh\"];\n\nexport function getSupportedThinkingLevels<TApi extends Api>(model: Model<TApi>): ModelThinkingLevel[] {\n\tif (!model.reasoning) return [\"off\"];\n\n\treturn EXTENDED_THINKING_LEVELS.filter((level) => {\n\t\tconst mapped = model.thinkingLevelMap?.[level];\n\t\tif (mapped === null) return false;\n\t\tif (level === \"xhigh\") return mapped !== undefined;\n\t\treturn true;\n\t});\n}\n\nexport function clampThinkingLevel<TApi extends Api>(\n\tmodel: Model<TApi>,\n\tlevel: ModelThinkingLevel,\n): ModelThinkingLevel {\n\tconst availableLevels = getSupportedThinkingLevels(model);\n\tif (availableLevels.includes(level)) return level;\n\n\tconst requestedIndex = EXTENDED_THINKING_LEVELS.indexOf(level);\n\tif (requestedIndex === -1) return availableLevels[0] ?? \"off\";\n\n\tfor (let i = requestedIndex; i < EXTENDED_THINKING_LEVELS.length; i++) {\n\t\tconst candidate = EXTENDED_THINKING_LEVELS[i];\n\t\tif (availableLevels.includes(candidate)) return candidate;\n\t}\n\tfor (let i = requestedIndex - 1; i >= 0; i--) {\n\t\tconst candidate = EXTENDED_THINKING_LEVELS[i];\n\t\tif (availableLevels.includes(candidate)) return candidate;\n\t}\n\treturn availableLevels[0] ?? \"off\";\n}\n\n/**\n * Check if two models are equal by comparing both their id and provider.\n * Returns false if either model is null or undefined.\n */\nexport function modelsAreEqual<TApi extends Api>(\n\ta: Model<TApi> | null | undefined,\n\tb: Model<TApi> | null | undefined,\n): boolean {\n\tif (!a || !b) return false;\n\treturn a.id === b.id && a.provider === b.provider;\n}\n"]}

package/dist/oauth.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"oauth.d.ts","sourceRoot":"","sources":["../src/oauth.ts"],"names":[],"mappings":"AAAA,cAAc,wBAAwB,CAAC"}
+{"version":3,"file":"oauth.d.ts","sourceRoot":"","sources":["../src/oauth.ts"],"names":[],"mappings":"AAAA,cAAc,wBAAwB,CAAC","sourcesContent":["export * from \"./utils/oauth/index.js\";\n"]}

package/dist/providers/amazon-bedrock.d.ts

@@ -1,4 +1,5 @@
 import type { SimpleStreamOptions, StreamFunction, StreamOptions, ThinkingBudgets, ThinkingLevel } from "../types.js";
+export type BedrockThinkingDisplay = "summarized" | "omitted";
 export interface BedrockOptions extends StreamOptions {
     region?: string;
     profile?: string;
@@ -9,6 +10,28 @@ export interface BedrockOptions extends StreamOptions {
     reasoning?: ThinkingLevel;
     thinkingBudgets?: ThinkingBudgets;
     interleavedThinking?: boolean;
+    /**
+     * Controls how Claude's thinking content is returned in responses.
+     * - "summarized": Thinking blocks contain summarized thinking text (default here).
+     * - "omitted": Thinking content is redacted but the signature still travels back
+     *   for multi-turn continuity, reducing time-to-first-text-token.
+     *
+     * Note: Anthropic's API default for Claude Opus 4.7 and Mythos Preview is
+     * "omitted". We default to "summarized" here to keep behavior consistent with
+     * older Claude 4 models. Only applies to Claude models on Bedrock.
+     */
+    thinkingDisplay?: BedrockThinkingDisplay;
+    /** Key-value pairs attached to the inference request for cost allocation tagging.
+     * Keys: max 64 chars, no `aws:` prefix. Values: max 256 chars. Max 50 pairs.
+     * Tags appear in AWS Cost Explorer split cost allocation data.
+     * @see https://docs.aws.amazon.com/bedrock/latest/APIReference/API_runtime_ConverseStream.html */
+    requestMetadata?: Record<string, string>;
+    /** Bearer token for Bedrock API key authentication.
+     * When set, bypasses SigV4 signing and sends Authorization: Bearer <token> instead.
+     * Requires `bedrock:CallWithBearerToken` IAM permission on the token's identity.
+     * Set via AWS_BEARER_TOKEN_BEDROCK env var or pass directly.
+     * @see https://docs.aws.amazon.com/service-authorization/latest/reference/list_amazonbedrock.html */
+    bearerToken?: string;
 }
 export declare const streamBedrock: StreamFunction<"bedrock-converse-stream", BedrockOptions>;
 export declare const streamSimpleBedrock: StreamFunction<"bedrock-converse-stream", SimpleStreamOptions>;

package/dist/providers/amazon-bedrock.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"amazon-bedrock.d.ts","sourceRoot":"","sources":["../../src/providers/amazon-bedrock.ts"],"names":[],"mappings":"AAuBA,OAAO,KAAK,EAMX,mBAAmB,EAEnB,cAAc,EACd,aAAa,EAEb,eAAe,EAEf,aAAa,EAIb,MAAM,aAAa,CAAC;AAOrB,MAAM,WAAW,cAAe,SAAQ,aAAa;IACpD,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,UAAU,CAAC,EAAE,MAAM,GAAG,KAAK,GAAG,MAAM,GAAG;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,MAAM,CAAA;KAAE,CAAC;IAEtE,SAAS,CAAC,EAAE,aAAa,CAAC;IAE1B,eAAe,CAAC,EAAE,eAAe,CAAC;IAElC,mBAAmB,CAAC,EAAE,OAAO,CAAC;CAC9B;AAID,eAAO,MAAM,aAAa,EAAE,cAAc,CAAC,yBAAyB,EAAE,cAAc,CAwJnF,CAAC;AAEF,eAAO,MAAM,mBAAmB,EAAE,cAAc,CAAC,yBAAyB,EAAE,mBAAmB,CA0C9F,CAAC"}
+{"version":3,"file":"amazon-bedrock.d.ts","sourceRoot":"","sources":["../../src/providers/amazon-bedrock.ts"],"names":[],"mappings":"AAwBA,OAAO,KAAK,EAMX,mBAAmB,EAEnB,cAAc,EACd,aAAa,EAEb,eAAe,EAEf,aAAa,EAIb,MAAM,aAAa,CAAC;AAOrB,MAAM,MAAM,sBAAsB,GAAG,YAAY,GAAG,SAAS,CAAC;AAE9D,MAAM,WAAW,cAAe,SAAQ,aAAa;IACpD,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,UAAU,CAAC,EAAE,MAAM,GAAG,KAAK,GAAG,MAAM,GAAG;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,MAAM,CAAA;KAAE,CAAC;IAEtE,SAAS,CAAC,EAAE,aAAa,CAAC;IAE1B,eAAe,CAAC,EAAE,eAAe,CAAC;IAElC,mBAAmB,CAAC,EAAE,OAAO,CAAC;IAC9B;;;;;;;;;OASG;IACH,eAAe,CAAC,EAAE,sBAAsB,CAAC;IACzC;;;sGAGkG;IAClG,eAAe,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACzC;;;;yGAIqG;IACrG,WAAW,CAAC,EAAE,MAAM,CAAC;CACrB;AAID,eAAO,MAAM,aAAa,EAAE,cAAc,CAAC,yBAAyB,EAAE,cAAc,CAgMnF,CAAC;AAgCF,eAAO,MAAM,mBAAmB,EAAE,cAAc,CAAC,yBAAyB,EAAE,mBAAmB,CA0C9F,CAAC","sourcesContent":["import {\n\tBedrockRuntimeClient,\n\ttype BedrockRuntimeClientConfig,\n\tBedrockRuntimeServiceException,\n\tStopReason as BedrockStopReason,\n\ttype Tool as BedrockTool,\n\tCachePointType,\n\tCacheTTL,\n\ttype ContentBlock,\n\ttype ContentBlockDeltaEvent,\n\ttype ContentBlockStartEvent,\n\ttype ContentBlockStopEvent,\n\tConversationRole,\n\tConverseStreamCommand,\n\ttype ConverseStreamMetadataEvent,\n\tImageFormat,\n\ttype Message,\n\ttype SystemContentBlock,\n\ttype ToolChoice,\n\ttype ToolConfiguration,\n\tToolResultStatus,\n} from \"@aws-sdk/client-bedrock-runtime\";\nimport type { DocumentType } from \"@smithy/types\";\nimport { calculateCost } from \"../models.js\";\nimport type {\n\tApi,\n\tAssistantMessage,\n\tCacheRetention,\n\tContext,\n\tModel,\n\tSimpleStreamOptions,\n\tStopReason,\n\tStreamFunction,\n\tStreamOptions,\n\tTextContent,\n\tThinkingBudgets,\n\tThinkingContent,\n\tThinkingLevel,\n\tTool,\n\tToolCall,\n\tToolResultMessage,\n} from \"../types.js\";\nimport { AssistantMessageEventStream } from \"../utils/event-stream.js\";\nimport { parseStreamingJson } from \"../utils/json-parse.js\";\nimport { sanitizeSurrogates } from \"../utils/sanitize-unicode.js\";\nimport { adjustMaxTokensForThinking, buildBaseOptions, clampReasoning } from \"./simple-options.js\";\nimport { transformMessages } from \"./transform-messages.js\";\n\nexport type BedrockThinkingDisplay = \"summarized\" | \"omitted\";\n\nexport interface BedrockOptions extends StreamOptions {\n\tregion?: string;\n\tprofile?: string;\n\ttoolChoice?: \"auto\" | \"any\" | \"none\" | { type: \"tool\"; name: string };\n\t/* See https://docs.aws.amazon.com/bedrock/latest/userguide/inference-reasoning.html for supported models. */\n\treasoning?: ThinkingLevel;\n\t/* Custom token budgets per thinking level. Overrides default budgets. */\n\tthinkingBudgets?: ThinkingBudgets;\n\t/* Only supported by Claude 4.x models, see https://docs.aws.amazon.com/bedrock/latest/userguide/claude-messages-extended-thinking.html#claude-messages-extended-thinking-tool-use-interleaved */\n\tinterleavedThinking?: boolean;\n\t/**\n\t * Controls how Claude's thinking content is returned in responses.\n\t * - \"summarized\": Thinking blocks contain summarized thinking text (default here).\n\t * - \"omitted\": Thinking content is redacted but the signature still travels back\n\t *   for multi-turn continuity, reducing time-to-first-text-token.\n\t *\n\t * Note: Anthropic's API default for Claude Opus 4.7 and Mythos Preview is\n\t * \"omitted\". We default to \"summarized\" here to keep behavior consistent with\n\t * older Claude 4 models. Only applies to Claude models on Bedrock.\n\t */\n\tthinkingDisplay?: BedrockThinkingDisplay;\n\t/** Key-value pairs attached to the inference request for cost allocation tagging.\n\t * Keys: max 64 chars, no `aws:` prefix. Values: max 256 chars. Max 50 pairs.\n\t * Tags appear in AWS Cost Explorer split cost allocation data.\n\t * @see https://docs.aws.amazon.com/bedrock/latest/APIReference/API_runtime_ConverseStream.html */\n\trequestMetadata?: Record<string, string>;\n\t/** Bearer token for Bedrock API key authentication.\n\t * When set, bypasses SigV4 signing and sends Authorization: Bearer <token> instead.\n\t * Requires `bedrock:CallWithBearerToken` IAM permission on the token's identity.\n\t * Set via AWS_BEARER_TOKEN_BEDROCK env var or pass directly.\n\t * @see https://docs.aws.amazon.com/service-authorization/latest/reference/list_amazonbedrock.html */\n\tbearerToken?: string;\n}\n\ntype Block = (TextContent | ThinkingContent | ToolCall) & { index?: number; partialJson?: string };\n\nexport const streamBedrock: StreamFunction<\"bedrock-converse-stream\", BedrockOptions> = (\n\tmodel: Model<\"bedrock-converse-stream\">,\n\tcontext: Context,\n\toptions: BedrockOptions = {},\n): AssistantMessageEventStream => {\n\tconst stream = new AssistantMessageEventStream();\n\n\t(async () => {\n\t\tconst output: AssistantMessage = {\n\t\t\trole: \"assistant\",\n\t\t\tcontent: [],\n\t\t\tapi: \"bedrock-converse-stream\" as Api,\n\t\t\tprovider: model.provider,\n\t\t\tmodel: model.id,\n\t\t\tusage: {\n\t\t\t\tinput: 0,\n\t\t\t\toutput: 0,\n\t\t\t\tcacheRead: 0,\n\t\t\t\tcacheWrite: 0,\n\t\t\t\ttotalTokens: 0,\n\t\t\t\tcost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0, total: 0 },\n\t\t\t},\n\t\t\tstopReason: \"stop\",\n\t\t\ttimestamp: Date.now(),\n\t\t};\n\n\t\tconst blocks = output.content as Block[];\n\n\t\tconst config: BedrockRuntimeClientConfig = {\n\t\t\tprofile: options.profile,\n\t\t};\n\t\tconst configuredRegion = getConfiguredBedrockRegion(options);\n\t\tconst hasConfiguredProfile = hasConfiguredBedrockProfile();\n\t\tconst endpointRegion = getStandardBedrockEndpointRegion(model.baseUrl);\n\t\tconst useExplicitEndpoint = shouldUseExplicitBedrockEndpoint(\n\t\t\tmodel.baseUrl,\n\t\t\tconfiguredRegion,\n\t\t\thasConfiguredProfile,\n\t\t);\n\n\t\t// Only pin standard AWS Bedrock runtime endpoints when no region/profile is configured.\n\t\t// This preserves custom endpoints (VPC/proxy) from #3402 without forcing built-in\n\t\t// catalog defaults such as us-east-1 to override AWS_REGION/AWS_PROFILE.\n\t\tif (useExplicitEndpoint) {\n\t\t\tconfig.endpoint = model.baseUrl;\n\t\t}\n\n\t\t// Resolve bearer token for Bedrock API key auth.\n\t\tconst bearerToken = options.bearerToken || process.env.AWS_BEARER_TOKEN_BEDROCK || undefined;\n\t\tconst useBearerToken = bearerToken !== undefined && process.env.AWS_BEDROCK_SKIP_AUTH !== \"1\";\n\n\t\t// in Node.js/Bun environment only\n\t\tif (typeof process !== \"undefined\" && (process.versions?.node || process.versions?.bun)) {\n\t\t\t// Region resolution: explicit option > env vars > SDK default chain.\n\t\t\t// When AWS_PROFILE is set, we leave region undefined so the SDK can\n\t\t\t// resovle it from aws profile configs. Otherwise fall back to us-east-1.\n\t\t\tif (configuredRegion) {\n\t\t\t\tconfig.region = configuredRegion;\n\t\t\t} else if (endpointRegion && useExplicitEndpoint) {\n\t\t\t\tconfig.region = endpointRegion;\n\t\t\t} else if (!hasConfiguredProfile) {\n\t\t\t\tconfig.region = \"us-east-1\";\n\t\t\t}\n\n\t\t\t// Support proxies that don't need authentication\n\t\t\tif (process.env.AWS_BEDROCK_SKIP_AUTH === \"1\") {\n\t\t\t\tconfig.credentials = {\n\t\t\t\t\taccessKeyId: \"dummy-access-key\",\n\t\t\t\t\tsecretAccessKey: \"dummy-secret-key\",\n\t\t\t\t};\n\t\t\t}\n\n\t\t\tif (\n\t\t\t\tprocess.env.HTTP_PROXY ||\n\t\t\t\tprocess.env.HTTPS_PROXY ||\n\t\t\t\tprocess.env.NO_PROXY ||\n\t\t\t\tprocess.env.http_proxy ||\n\t\t\t\tprocess.env.https_proxy ||\n\t\t\t\tprocess.env.no_proxy\n\t\t\t) {\n\t\t\t\tconst nodeHttpHandler = await import(\"@smithy/node-http-handler\");\n\t\t\t\tconst proxyAgent = await import(\"proxy-agent\");\n\n\t\t\t\tconst agent = new proxyAgent.ProxyAgent();\n\n\t\t\t\t// Bedrock runtime uses NodeHttp2Handler by default since v3.798.0, which is based\n\t\t\t\t// on `http2` module and has no support for http agent.\n\t\t\t\t// Use NodeHttpHandler to support http agent.\n\t\t\t\tconfig.requestHandler = new nodeHttpHandler.NodeHttpHandler({\n\t\t\t\t\thttpAgent: agent,\n\t\t\t\t\thttpsAgent: agent,\n\t\t\t\t});\n\t\t\t} else if (process.env.AWS_BEDROCK_FORCE_HTTP1 === \"1\") {\n\t\t\t\t// Some custom endpoints require HTTP/1.1 instead of HTTP/2\n\t\t\t\tconst nodeHttpHandler = await import(\"@smithy/node-http-handler\");\n\t\t\t\tconfig.requestHandler = new nodeHttpHandler.NodeHttpHandler();\n\t\t\t}\n\t\t} else {\n\t\t\t// Non-Node environment (browser): fall back to us-east-1 since\n\t\t\t// there's no config file resolution available.\n\t\t\tconfig.region =\n\t\t\t\tconfiguredRegion || (endpointRegion && useExplicitEndpoint ? endpointRegion : undefined) || \"us-east-1\";\n\t\t}\n\n\t\tif (useBearerToken) {\n\t\t\tconfig.token = { token: bearerToken };\n\t\t\tconfig.authSchemePreference = [\"httpBearerAuth\"];\n\t\t}\n\n\t\ttry {\n\t\t\tconst client = new BedrockRuntimeClient(config);\n\t\t\tconst cacheRetention = resolveCacheRetention(options.cacheRetention);\n\t\t\tlet commandInput = {\n\t\t\t\tmodelId: model.id,\n\t\t\t\tmessages: convertMessages(context, model, cacheRetention),\n\t\t\t\tsystem: buildSystemPrompt(context.systemPrompt, model, cacheRetention),\n\t\t\t\tinferenceConfig: {\n\t\t\t\t\t...(options.maxTokens !== undefined && { maxTokens: options.maxTokens }),\n\t\t\t\t\t...(options.temperature !== undefined && { temperature: options.temperature }),\n\t\t\t\t},\n\t\t\t\ttoolConfig: convertToolConfig(context.tools, options.toolChoice),\n\t\t\t\tadditionalModelRequestFields: buildAdditionalModelRequestFields(model, options),\n\t\t\t\t...(options.requestMetadata !== undefined && { requestMetadata: options.requestMetadata }),\n\t\t\t};\n\t\t\tconst nextCommandInput = await options?.onPayload?.(commandInput, model);\n\t\t\tif (nextCommandInput !== undefined) {\n\t\t\t\tcommandInput = nextCommandInput as typeof commandInput;\n\t\t\t}\n\t\t\tconst command = new ConverseStreamCommand(commandInput);\n\n\t\t\tconst response = await client.send(command, { abortSignal: options.signal });\n\t\t\tif (response.$metadata.httpStatusCode !== undefined) {\n\t\t\t\tconst responseHeaders: Record<string, string> = {};\n\t\t\t\tif (response.$metadata.requestId) {\n\t\t\t\t\tresponseHeaders[\"x-amzn-requestid\"] = response.$metadata.requestId;\n\t\t\t\t}\n\t\t\t\tawait options?.onResponse?.({ status: response.$metadata.httpStatusCode, headers: responseHeaders }, model);\n\t\t\t}\n\n\t\t\tfor await (const item of response.stream!) {\n\t\t\t\tif (item.messageStart) {\n\t\t\t\t\tif (item.messageStart.role !== ConversationRole.ASSISTANT) {\n\t\t\t\t\t\tthrow new Error(\"Unexpected assistant message start but got user message start instead\");\n\t\t\t\t\t}\n\t\t\t\t\tstream.push({ type: \"start\", partial: output });\n\t\t\t\t} else if (item.contentBlockStart) {\n\t\t\t\t\thandleContentBlockStart(item.contentBlockStart, blocks, output, stream);\n\t\t\t\t} else if (item.contentBlockDelta) {\n\t\t\t\t\thandleContentBlockDelta(item.contentBlockDelta, blocks, output, stream);\n\t\t\t\t} else if (item.contentBlockStop) {\n\t\t\t\t\thandleContentBlockStop(item.contentBlockStop, blocks, output, stream);\n\t\t\t\t} else if (item.messageStop) {\n\t\t\t\t\toutput.stopReason = mapStopReason(item.messageStop.stopReason);\n\t\t\t\t} else if (item.metadata) {\n\t\t\t\t\thandleMetadata(item.metadata, model, output);\n\t\t\t\t} else if (item.internalServerException) {\n\t\t\t\t\tthrow item.internalServerException;\n\t\t\t\t} else if (item.modelStreamErrorException) {\n\t\t\t\t\tthrow item.modelStreamErrorException;\n\t\t\t\t} else if (item.validationException) {\n\t\t\t\t\tthrow item.validationException;\n\t\t\t\t} else if (item.throttlingException) {\n\t\t\t\t\tthrow item.throttlingException;\n\t\t\t\t} else if (item.serviceUnavailableException) {\n\t\t\t\t\tthrow item.serviceUnavailableException;\n\t\t\t\t}\n\t\t\t}\n\n\t\t\tif (options.signal?.aborted) {\n\t\t\t\tthrow new Error(\"Request was aborted\");\n\t\t\t}\n\n\t\t\tif (output.stopReason === \"error\" || output.stopReason === \"aborted\") {\n\t\t\t\tthrow new Error(\"An unknown error occurred\");\n\t\t\t}\n\n\t\t\tstream.push({ type: \"done\", reason: output.stopReason, message: output });\n\t\t\tstream.end();\n\t\t} catch (error) {\n\t\t\tfor (const block of output.content) {\n\t\t\t\tdelete (block as Block).index;\n\t\t\t\t// partialJson is only a streaming scratch buffer; never persist it.\n\t\t\t\tdelete (block as Block).partialJson;\n\t\t\t}\n\t\t\toutput.stopReason = options.signal?.aborted ? \"aborted\" : \"error\";\n\t\t\toutput.errorMessage = formatBedrockError(error);\n\t\t\tstream.push({ type: \"error\", reason: output.stopReason, error: output });\n\t\t\tstream.end();\n\t\t}\n\t})();\n\n\treturn stream;\n};\n\n/**\n * Human-readable prefixes for Bedrock SDK exception names.\n * The downstream retry logic in agent-session matches patterns like\n * `server.?error` and `service.?unavailable`, so we preserve the legacy\n * prefix format rather than using the raw SDK exception name.\n */\nconst BEDROCK_ERROR_PREFIXES: Record<string, string> = {\n\tInternalServerException: \"Internal server error\",\n\tModelStreamErrorException: \"Model stream error\",\n\tValidationException: \"Validation error\",\n\tThrottlingException: \"Throttling error\",\n\tServiceUnavailableException: \"Service unavailable\",\n};\n\n/**\n * Format a Bedrock error with a human-readable prefix.\n * AWS SDK exceptions (both from `client.send()` and from stream event items)\n * extend BedrockRuntimeServiceException. We map the `.name` to a stable\n * human-readable prefix so downstream consumers (retry logic, context-overflow\n * detection) can distinguish error categories via simple string matching.\n */\nfunction formatBedrockError(error: unknown): string {\n\tconst message = error instanceof Error ? error.message : JSON.stringify(error);\n\tif (error instanceof BedrockRuntimeServiceException) {\n\t\tconst prefix = BEDROCK_ERROR_PREFIXES[error.name] ?? error.name;\n\t\treturn `${prefix}: ${message}`;\n\t}\n\treturn message;\n}\n\nexport const streamSimpleBedrock: StreamFunction<\"bedrock-converse-stream\", SimpleStreamOptions> = (\n\tmodel: Model<\"bedrock-converse-stream\">,\n\tcontext: Context,\n\toptions?: SimpleStreamOptions,\n): AssistantMessageEventStream => {\n\tconst base = buildBaseOptions(model, options, undefined);\n\tif (!options?.reasoning) {\n\t\treturn streamBedrock(model, context, { ...base, reasoning: undefined } satisfies BedrockOptions);\n\t}\n\n\tif (isAnthropicClaudeModel(model)) {\n\t\tif (supportsAdaptiveThinking(model.id, model.name)) {\n\t\t\treturn streamBedrock(model, context, {\n\t\t\t\t...base,\n\t\t\t\treasoning: options.reasoning,\n\t\t\t\tthinkingBudgets: options.thinkingBudgets,\n\t\t\t} satisfies BedrockOptions);\n\t\t}\n\n\t\tconst adjusted = adjustMaxTokensForThinking(\n\t\t\tbase.maxTokens || 0,\n\t\t\tmodel.maxTokens,\n\t\t\toptions.reasoning,\n\t\t\toptions.thinkingBudgets,\n\t\t);\n\n\t\treturn streamBedrock(model, context, {\n\t\t\t...base,\n\t\t\tmaxTokens: adjusted.maxTokens,\n\t\t\treasoning: options.reasoning,\n\t\t\tthinkingBudgets: {\n\t\t\t\t...(options.thinkingBudgets || {}),\n\t\t\t\t[clampReasoning(options.reasoning)!]: adjusted.thinkingBudget,\n\t\t\t},\n\t\t} satisfies BedrockOptions);\n\t}\n\n\treturn streamBedrock(model, context, {\n\t\t...base,\n\t\treasoning: options.reasoning,\n\t\tthinkingBudgets: options.thinkingBudgets,\n\t} satisfies BedrockOptions);\n};\n\nfunction handleContentBlockStart(\n\tevent: ContentBlockStartEvent,\n\tblocks: Block[],\n\toutput: AssistantMessage,\n\tstream: AssistantMessageEventStream,\n): void {\n\tconst index = event.contentBlockIndex!;\n\tconst start = event.start;\n\n\tif (start?.toolUse) {\n\t\tconst block: Block = {\n\t\t\ttype: \"toolCall\",\n\t\t\tid: start.toolUse.toolUseId || \"\",\n\t\t\tname: start.toolUse.name || \"\",\n\t\t\targuments: {},\n\t\t\tpartialJson: \"\",\n\t\t\tindex,\n\t\t};\n\t\toutput.content.push(block);\n\t\tstream.push({ type: \"toolcall_start\", contentIndex: blocks.length - 1, partial: output });\n\t}\n}\n\nfunction handleContentBlockDelta(\n\tevent: ContentBlockDeltaEvent,\n\tblocks: Block[],\n\toutput: AssistantMessage,\n\tstream: AssistantMessageEventStream,\n): void {\n\tconst contentBlockIndex = event.contentBlockIndex!;\n\tconst delta = event.delta;\n\tlet index = blocks.findIndex((b) => b.index === contentBlockIndex);\n\tlet block = blocks[index];\n\n\tif (delta?.text !== undefined) {\n\t\t// If no text block exists yet, create one, as `handleContentBlockStart` is not sent for text blocks\n\t\tif (!block) {\n\t\t\tconst newBlock: Block = { type: \"text\", text: \"\", index: contentBlockIndex };\n\t\t\toutput.content.push(newBlock);\n\t\t\tindex = blocks.length - 1;\n\t\t\tblock = blocks[index];\n\t\t\tstream.push({ type: \"text_start\", contentIndex: index, partial: output });\n\t\t}\n\t\tif (block.type === \"text\") {\n\t\t\tblock.text += delta.text;\n\t\t\tstream.push({ type: \"text_delta\", contentIndex: index, delta: delta.text, partial: output });\n\t\t}\n\t} else if (delta?.toolUse && block?.type === \"toolCall\") {\n\t\tblock.partialJson = (block.partialJson || \"\") + (delta.toolUse.input || \"\");\n\t\tblock.arguments = parseStreamingJson(block.partialJson);\n\t\tstream.push({ type: \"toolcall_delta\", contentIndex: index, delta: delta.toolUse.input || \"\", partial: output });\n\t} else if (delta?.reasoningContent) {\n\t\tlet thinkingBlock = block;\n\t\tlet thinkingIndex = index;\n\n\t\tif (!thinkingBlock) {\n\t\t\tconst newBlock: Block = { type: \"thinking\", thinking: \"\", thinkingSignature: \"\", index: contentBlockIndex };\n\t\t\toutput.content.push(newBlock);\n\t\t\tthinkingIndex = blocks.length - 1;\n\t\t\tthinkingBlock = blocks[thinkingIndex];\n\t\t\tstream.push({ type: \"thinking_start\", contentIndex: thinkingIndex, partial: output });\n\t\t}\n\n\t\tif (thinkingBlock?.type === \"thinking\") {\n\t\t\tif (delta.reasoningContent.text) {\n\t\t\t\tthinkingBlock.thinking += delta.reasoningContent.text;\n\t\t\t\tstream.push({\n\t\t\t\t\ttype: \"thinking_delta\",\n\t\t\t\t\tcontentIndex: thinkingIndex,\n\t\t\t\t\tdelta: delta.reasoningContent.text,\n\t\t\t\t\tpartial: output,\n\t\t\t\t});\n\t\t\t}\n\t\t\tif (delta.reasoningContent.signature) {\n\t\t\t\tthinkingBlock.thinkingSignature =\n\t\t\t\t\t(thinkingBlock.thinkingSignature || \"\") + delta.reasoningContent.signature;\n\t\t\t}\n\t\t}\n\t}\n}\n\nfunction handleMetadata(\n\tevent: ConverseStreamMetadataEvent,\n\tmodel: Model<\"bedrock-converse-stream\">,\n\toutput: AssistantMessage,\n): void {\n\tif (event.usage) {\n\t\toutput.usage.input = event.usage.inputTokens || 0;\n\t\toutput.usage.output = event.usage.outputTokens || 0;\n\t\toutput.usage.cacheRead = event.usage.cacheReadInputTokens || 0;\n\t\toutput.usage.cacheWrite = event.usage.cacheWriteInputTokens || 0;\n\t\toutput.usage.totalTokens = event.usage.totalTokens || output.usage.input + output.usage.output;\n\t\tcalculateCost(model, output.usage);\n\t}\n}\n\nfunction handleContentBlockStop(\n\tevent: ContentBlockStopEvent,\n\tblocks: Block[],\n\toutput: AssistantMessage,\n\tstream: AssistantMessageEventStream,\n): void {\n\tconst index = blocks.findIndex((b) => b.index === event.contentBlockIndex);\n\tconst block = blocks[index];\n\tif (!block) return;\n\tdelete (block as Block).index;\n\n\tswitch (block.type) {\n\t\tcase \"text\":\n\t\t\tstream.push({ type: \"text_end\", contentIndex: index, content: block.text, partial: output });\n\t\t\tbreak;\n\t\tcase \"thinking\":\n\t\t\tstream.push({ type: \"thinking_end\", contentIndex: index, content: block.thinking, partial: output });\n\t\t\tbreak;\n\t\tcase \"toolCall\":\n\t\t\tblock.arguments = parseStreamingJson(block.partialJson);\n\t\t\t// Finalize in-place and strip the scratch buffer so replay only\n\t\t\t// carries parsed arguments.\n\t\t\tdelete (block as Block).partialJson;\n\t\t\tstream.push({ type: \"toolcall_end\", contentIndex: index, toolCall: block, partial: output });\n\t\t\tbreak;\n\t}\n}\n\n/**\n * Check if the model supports adaptive thinking (Opus 4.6+, Sonnet 4.6).\n * Checks both model ID and model name to support application inference profiles\n * whose ARNs don't contain the model name.\n */\nfunction getModelMatchCandidates(modelId: string, modelName?: string): string[] {\n\tconst values = modelName ? [modelId, modelName] : [modelId];\n\treturn values.flatMap((value) => {\n\t\tconst lower = value.toLowerCase();\n\t\treturn [lower, lower.replace(/[\\s_.:]+/g, \"-\")];\n\t});\n}\n\nfunction supportsAdaptiveThinking(modelId: string, modelName?: string): boolean {\n\tconst candidates = getModelMatchCandidates(modelId, modelName);\n\treturn candidates.some((s) => s.includes(\"opus-4-6\") || s.includes(\"opus-4-7\") || s.includes(\"sonnet-4-6\"));\n}\n\nfunction supportsNativeXhighEffort(model: Model<\"bedrock-converse-stream\">): boolean {\n\tconst candidates = getModelMatchCandidates(model.id, model.name);\n\treturn candidates.some((s) => s.includes(\"opus-4-7\"));\n}\n\nfunction mapThinkingLevelToEffort(\n\tmodel: Model<\"bedrock-converse-stream\">,\n\tlevel: SimpleStreamOptions[\"reasoning\"],\n): \"low\" | \"medium\" | \"high\" | \"xhigh\" | \"max\" {\n\tif (level === \"xhigh\" && supportsNativeXhighEffort(model)) return \"xhigh\";\n\n\tconst mapped = level ? model.thinkingLevelMap?.[level] : undefined;\n\tif (typeof mapped === \"string\") return mapped as \"low\" | \"medium\" | \"high\" | \"xhigh\" | \"max\";\n\n\tswitch (level) {\n\t\tcase \"minimal\":\n\t\tcase \"low\":\n\t\t\treturn \"low\";\n\t\tcase \"medium\":\n\t\t\treturn \"medium\";\n\t\tcase \"high\":\n\t\t\treturn \"high\";\n\t\tdefault:\n\t\t\treturn \"high\";\n\t}\n}\n\n/**\n * Resolve cache retention preference.\n * Defaults to \"short\" and uses PI_CACHE_RETENTION for backward compatibility.\n */\nfunction resolveCacheRetention(cacheRetention?: CacheRetention): CacheRetention {\n\tif (cacheRetention) {\n\t\treturn cacheRetention;\n\t}\n\tif (typeof process !== \"undefined\" && process.env.PI_CACHE_RETENTION === \"long\") {\n\t\treturn \"long\";\n\t}\n\treturn \"short\";\n}\n\n/**\n * Check if the model is an Anthropic Claude model on Bedrock.\n * Checks both model ID and model name to support application inference profiles\n * whose ARNs don't contain the model name.\n */\nfunction isAnthropicClaudeModel(model: Model<\"bedrock-converse-stream\">): boolean {\n\tconst id = model.id.toLowerCase();\n\tconst name = model.name?.toLowerCase() ?? \"\";\n\treturn (\n\t\tid.includes(\"anthropic.claude\") ||\n\t\tid.includes(\"anthropic/claude\") ||\n\t\tname.includes(\"anthropic.claude\") ||\n\t\tname.includes(\"anthropic/claude\") ||\n\t\tname.includes(\"claude\")\n\t);\n}\n\n/**\n * Check if the model supports prompt caching.\n * Supported: Claude 3.5 Haiku, Claude 3.7 Sonnet, Claude 4.x models\n *\n * For base models and system-defined inference profiles the model ID / ARN\n * contains the model name, so we can decide locally.\n *\n * For application inference profiles (whose ARNs don't contain the model name),\n * also checks model.name which is user-controlled via models.json or registerProvider.\n * As a last resort, set AWS_BEDROCK_FORCE_CACHE=1 to enable cache points.\n * Amazon Nova models have automatic caching and don't need explicit cache points.\n */\nfunction supportsPromptCaching(model: Model<\"bedrock-converse-stream\">): boolean {\n\tconst candidates = getModelMatchCandidates(model.id, model.name);\n\n\tconst hasClaudeRef = candidates.some((s) => s.includes(\"claude\"));\n\tif (!hasClaudeRef) {\n\t\t// Application inference profiles don't contain the model name in the ARN.\n\t\t// Allow users to force cache points via environment variable.\n\t\tif (typeof process !== \"undefined\" && process.env.AWS_BEDROCK_FORCE_CACHE === \"1\") return true;\n\t\treturn false;\n\t}\n\t// Claude 4.x models (opus-4, sonnet-4, haiku-4)\n\tif (candidates.some((s) => s.includes(\"-4-\"))) return true;\n\t// Claude 3.7 Sonnet\n\tif (candidates.some((s) => s.includes(\"claude-3-7-sonnet\"))) return true;\n\t// Claude 3.5 Haiku\n\tif (candidates.some((s) => s.includes(\"claude-3-5-haiku\"))) return true;\n\treturn false;\n}\n\n/**\n * Check if the model supports thinking signatures in reasoningContent.\n * Only Anthropic Claude models support the signature field.\n * Other models (OpenAI, Qwen, Minimax, Moonshot, etc.) reject it with:\n * \"This model doesn't support the reasoningContent.reasoningText.signature field\"\n *\n * Checks both model ID and model name to support application inference profiles.\n */\nfunction supportsThinkingSignature(model: Model<\"bedrock-converse-stream\">): boolean {\n\treturn isAnthropicClaudeModel(model);\n}\n\nfunction buildSystemPrompt(\n\tsystemPrompt: string | undefined,\n\tmodel: Model<\"bedrock-converse-stream\">,\n\tcacheRetention: CacheRetention,\n): SystemContentBlock[] | undefined {\n\tif (!systemPrompt) return undefined;\n\n\tconst blocks: SystemContentBlock[] = [{ text: sanitizeSurrogates(systemPrompt) }];\n\n\t// Add cache point for supported Claude models when caching is enabled\n\tif (cacheRetention !== \"none\" && supportsPromptCaching(model)) {\n\t\tblocks.push({\n\t\t\tcachePoint: { type: CachePointType.DEFAULT, ...(cacheRetention === \"long\" ? { ttl: CacheTTL.ONE_HOUR } : {}) },\n\t\t});\n\t}\n\n\treturn blocks;\n}\n\nfunction normalizeToolCallId(id: string): string {\n\tconst sanitized = id.replace(/[^a-zA-Z0-9_-]/g, \"_\");\n\treturn sanitized.length > 64 ? sanitized.slice(0, 64) : sanitized;\n}\n\nfunction convertMessages(\n\tcontext: Context,\n\tmodel: Model<\"bedrock-converse-stream\">,\n\tcacheRetention: CacheRetention,\n): Message[] {\n\tconst result: Message[] = [];\n\tconst transformedMessages = transformMessages(context.messages, model, normalizeToolCallId);\n\n\tfor (let i = 0; i < transformedMessages.length; i++) {\n\t\tconst m = transformedMessages[i];\n\n\t\tswitch (m.role) {\n\t\t\tcase \"user\":\n\t\t\t\tresult.push({\n\t\t\t\t\trole: ConversationRole.USER,\n\t\t\t\t\tcontent:\n\t\t\t\t\t\ttypeof m.content === \"string\"\n\t\t\t\t\t\t\t? [{ text: sanitizeSurrogates(m.content) }]\n\t\t\t\t\t\t\t: m.content.map((c) => {\n\t\t\t\t\t\t\t\t\tswitch (c.type) {\n\t\t\t\t\t\t\t\t\t\tcase \"text\":\n\t\t\t\t\t\t\t\t\t\t\treturn { text: sanitizeSurrogates(c.text) };\n\t\t\t\t\t\t\t\t\t\tcase \"image\":\n\t\t\t\t\t\t\t\t\t\t\treturn { image: createImageBlock(c.mimeType, c.data) };\n\t\t\t\t\t\t\t\t\t\tdefault:\n\t\t\t\t\t\t\t\t\t\t\tthrow new Error(\"Unknown user content type\");\n\t\t\t\t\t\t\t\t\t}\n\t\t\t\t\t\t\t\t}),\n\t\t\t\t});\n\t\t\t\tbreak;\n\t\t\tcase \"assistant\": {\n\t\t\t\t// Skip assistant messages with empty content (e.g., from aborted requests)\n\t\t\t\t// Bedrock rejects messages with empty content arrays\n\t\t\t\tif (m.content.length === 0) {\n\t\t\t\t\tcontinue;\n\t\t\t\t}\n\t\t\t\tconst contentBlocks: ContentBlock[] = [];\n\t\t\t\tfor (const c of m.content) {\n\t\t\t\t\tswitch (c.type) {\n\t\t\t\t\t\tcase \"text\":\n\t\t\t\t\t\t\t// Skip empty text blocks\n\t\t\t\t\t\t\tif (c.text.trim().length === 0) continue;\n\t\t\t\t\t\t\tcontentBlocks.push({ text: sanitizeSurrogates(c.text) });\n\t\t\t\t\t\t\tbreak;\n\t\t\t\t\t\tcase \"toolCall\":\n\t\t\t\t\t\t\tcontentBlocks.push({\n\t\t\t\t\t\t\t\ttoolUse: { toolUseId: c.id, name: c.name, input: c.arguments },\n\t\t\t\t\t\t\t});\n\t\t\t\t\t\t\tbreak;\n\t\t\t\t\t\tcase \"thinking\":\n\t\t\t\t\t\t\t// Skip empty thinking blocks\n\t\t\t\t\t\t\tif (c.thinking.trim().length === 0) continue;\n\t\t\t\t\t\t\t// Only Anthropic models support the signature field in reasoningText.\n\t\t\t\t\t\t\t// For other models, we omit the signature to avoid errors like:\n\t\t\t\t\t\t\t// \"This model doesn't support the reasoningContent.reasoningText.signature field\"\n\t\t\t\t\t\t\tif (supportsThinkingSignature(model)) {\n\t\t\t\t\t\t\t\t// Signatures arrive after thinking deltas. If a partial or externally\n\t\t\t\t\t\t\t\t// persisted message lacks a signature, Bedrock rejects the replayed\n\t\t\t\t\t\t\t\t// reasoning block. Fall back to plain text, matching Anthropic.\n\t\t\t\t\t\t\t\tif (!c.thinkingSignature || c.thinkingSignature.trim().length === 0) {\n\t\t\t\t\t\t\t\t\tcontentBlocks.push({ text: sanitizeSurrogates(c.thinking) });\n\t\t\t\t\t\t\t\t} else {\n\t\t\t\t\t\t\t\t\tcontentBlocks.push({\n\t\t\t\t\t\t\t\t\t\treasoningContent: {\n\t\t\t\t\t\t\t\t\t\t\treasoningText: {\n\t\t\t\t\t\t\t\t\t\t\t\ttext: sanitizeSurrogates(c.thinking),\n\t\t\t\t\t\t\t\t\t\t\t\tsignature: c.thinkingSignature,\n\t\t\t\t\t\t\t\t\t\t\t},\n\t\t\t\t\t\t\t\t\t\t},\n\t\t\t\t\t\t\t\t\t});\n\t\t\t\t\t\t\t\t}\n\t\t\t\t\t\t\t} else {\n\t\t\t\t\t\t\t\tcontentBlocks.push({\n\t\t\t\t\t\t\t\t\treasoningContent: {\n\t\t\t\t\t\t\t\t\t\treasoningText: { text: sanitizeSurrogates(c.thinking) },\n\t\t\t\t\t\t\t\t\t},\n\t\t\t\t\t\t\t\t});\n\t\t\t\t\t\t\t}\n\t\t\t\t\t\t\tbreak;\n\t\t\t\t\t\tdefault:\n\t\t\t\t\t\t\tthrow new Error(\"Unknown assistant content type\");\n\t\t\t\t\t}\n\t\t\t\t}\n\t\t\t\t// Skip if all content blocks were filtered out\n\t\t\t\tif (contentBlocks.length === 0) {\n\t\t\t\t\tcontinue;\n\t\t\t\t}\n\t\t\t\tresult.push({\n\t\t\t\t\trole: ConversationRole.ASSISTANT,\n\t\t\t\t\tcontent: contentBlocks,\n\t\t\t\t});\n\t\t\t\tbreak;\n\t\t\t}\n\t\t\tcase \"toolResult\": {\n\t\t\t\t// Collect all consecutive toolResult messages into a single user message\n\t\t\t\t// Bedrock requires all tool results to be in one message\n\t\t\t\tconst toolResults: ContentBlock.ToolResultMember[] = [];\n\n\t\t\t\t// Add current tool result with all content blocks combined\n\t\t\t\ttoolResults.push({\n\t\t\t\t\ttoolResult: {\n\t\t\t\t\t\ttoolUseId: m.toolCallId,\n\t\t\t\t\t\tcontent: m.content.map((c) =>\n\t\t\t\t\t\t\tc.type === \"image\"\n\t\t\t\t\t\t\t\t? { image: createImageBlock(c.mimeType, c.data) }\n\t\t\t\t\t\t\t\t: { text: sanitizeSurrogates(c.text) },\n\t\t\t\t\t\t),\n\t\t\t\t\t\tstatus: m.isError ? ToolResultStatus.ERROR : ToolResultStatus.SUCCESS,\n\t\t\t\t\t},\n\t\t\t\t});\n\n\t\t\t\t// Look ahead for consecutive toolResult messages\n\t\t\t\tlet j = i + 1;\n\t\t\t\twhile (j < transformedMessages.length && transformedMessages[j].role === \"toolResult\") {\n\t\t\t\t\tconst nextMsg = transformedMessages[j] as ToolResultMessage;\n\t\t\t\t\ttoolResults.push({\n\t\t\t\t\t\ttoolResult: {\n\t\t\t\t\t\t\ttoolUseId: nextMsg.toolCallId,\n\t\t\t\t\t\t\tcontent: nextMsg.content.map((c) =>\n\t\t\t\t\t\t\t\tc.type === \"image\"\n\t\t\t\t\t\t\t\t\t? { image: createImageBlock(c.mimeType, c.data) }\n\t\t\t\t\t\t\t\t\t: { text: sanitizeSurrogates(c.text) },\n\t\t\t\t\t\t\t),\n\t\t\t\t\t\t\tstatus: nextMsg.isError ? ToolResultStatus.ERROR : ToolResultStatus.SUCCESS,\n\t\t\t\t\t\t},\n\t\t\t\t\t});\n\t\t\t\t\tj++;\n\t\t\t\t}\n\n\t\t\t\t// Skip the messages we've already processed\n\t\t\t\ti = j - 1;\n\n\t\t\t\tresult.push({\n\t\t\t\t\trole: ConversationRole.USER,\n\t\t\t\t\tcontent: toolResults,\n\t\t\t\t});\n\t\t\t\tbreak;\n\t\t\t}\n\t\t\tdefault:\n\t\t\t\tthrow new Error(\"Unknown message role\");\n\t\t}\n\t}\n\n\t// Add cache point to the last user message for supported Claude models when caching is enabled\n\tif (cacheRetention !== \"none\" && supportsPromptCaching(model) && result.length > 0) {\n\t\tconst lastMessage = result[result.length - 1];\n\t\tif (lastMessage.role === ConversationRole.USER && lastMessage.content) {\n\t\t\t(lastMessage.content as ContentBlock[]).push({\n\t\t\t\tcachePoint: {\n\t\t\t\t\ttype: CachePointType.DEFAULT,\n\t\t\t\t\t...(cacheRetention === \"long\" ? { ttl: CacheTTL.ONE_HOUR } : {}),\n\t\t\t\t},\n\t\t\t});\n\t\t}\n\t}\n\n\treturn result;\n}\n\nfunction convertToolConfig(\n\ttools: Tool[] | undefined,\n\ttoolChoice: BedrockOptions[\"toolChoice\"],\n): ToolConfiguration | undefined {\n\tif (!tools?.length || toolChoice === \"none\") return undefined;\n\n\tconst bedrockTools: BedrockTool[] = tools.map((tool) => ({\n\t\ttoolSpec: {\n\t\t\tname: tool.name,\n\t\t\tdescription: tool.description,\n\t\t\tinputSchema: { json: tool.parameters as unknown as DocumentType },\n\t\t},\n\t}));\n\n\tlet bedrockToolChoice: ToolChoice | undefined;\n\tswitch (toolChoice) {\n\t\tcase \"auto\":\n\t\t\tbedrockToolChoice = { auto: {} };\n\t\t\tbreak;\n\t\tcase \"any\":\n\t\t\tbedrockToolChoice = { any: {} };\n\t\t\tbreak;\n\t\tdefault:\n\t\t\tif (toolChoice?.type === \"tool\") {\n\t\t\t\tbedrockToolChoice = { tool: { name: toolChoice.name } };\n\t\t\t}\n\t}\n\n\treturn { tools: bedrockTools, toolChoice: bedrockToolChoice };\n}\n\nfunction mapStopReason(reason: string | undefined): StopReason {\n\tswitch (reason) {\n\t\tcase BedrockStopReason.END_TURN:\n\t\tcase BedrockStopReason.STOP_SEQUENCE:\n\t\t\treturn \"stop\";\n\t\tcase BedrockStopReason.MAX_TOKENS:\n\t\tcase BedrockStopReason.MODEL_CONTEXT_WINDOW_EXCEEDED:\n\t\t\treturn \"length\";\n\t\tcase BedrockStopReason.TOOL_USE:\n\t\t\treturn \"toolUse\";\n\t\tdefault:\n\t\t\treturn \"error\";\n\t}\n}\n\nfunction getConfiguredBedrockRegion(options: BedrockOptions): string | undefined {\n\tif (typeof process === \"undefined\") {\n\t\treturn options.region;\n\t}\n\n\treturn options.region || process.env.AWS_REGION || process.env.AWS_DEFAULT_REGION || undefined;\n}\n\nfunction hasConfiguredBedrockProfile(): boolean {\n\tif (typeof process === \"undefined\") {\n\t\treturn false;\n\t}\n\n\treturn Boolean(process.env.AWS_PROFILE);\n}\n\nfunction getStandardBedrockEndpointRegion(baseUrl: string | undefined): string | undefined {\n\tif (!baseUrl) {\n\t\treturn undefined;\n\t}\n\n\ttry {\n\t\tconst { hostname } = new URL(baseUrl);\n\t\tconst match = hostname.toLowerCase().match(/^bedrock-runtime(?:-fips)?\\.([a-z0-9-]+)\\.amazonaws\\.com(?:\\.cn)?$/);\n\t\treturn match?.[1];\n\t} catch {\n\t\treturn undefined;\n\t}\n}\n\nfunction shouldUseExplicitBedrockEndpoint(\n\tbaseUrl: string,\n\tconfiguredRegion: string | undefined,\n\thasConfiguredProfile: boolean,\n): boolean {\n\tconst endpointRegion = getStandardBedrockEndpointRegion(baseUrl);\n\tif (!endpointRegion) {\n\t\treturn true;\n\t}\n\n\treturn !configuredRegion && !hasConfiguredProfile;\n}\n\nfunction isGovCloudBedrockTarget(model: Model<\"bedrock-converse-stream\">, options: BedrockOptions): boolean {\n\tconst region = getConfiguredBedrockRegion(options);\n\tif (region?.toLowerCase().startsWith(\"us-gov-\")) {\n\t\treturn true;\n\t}\n\n\tconst modelId = model.id.toLowerCase();\n\treturn modelId.startsWith(\"us-gov.\") || modelId.startsWith(\"arn:aws-us-gov:\");\n}\n\nfunction buildAdditionalModelRequestFields(\n\tmodel: Model<\"bedrock-converse-stream\">,\n\toptions: BedrockOptions,\n): Record<string, any> | undefined {\n\tif (!options.reasoning || !model.reasoning) {\n\t\treturn undefined;\n\t}\n\n\tif (isAnthropicClaudeModel(model)) {\n\t\t// GovCloud Bedrock currently rejects the Claude thinking.display field.\n\t\t// Omit it there until the GovCloud Converse schema catches up.\n\t\tconst display = isGovCloudBedrockTarget(model, options) ? undefined : (options.thinkingDisplay ?? \"summarized\");\n\t\tconst result: Record<string, any> = supportsAdaptiveThinking(model.id, model.name)\n\t\t\t? {\n\t\t\t\t\tthinking: { type: \"adaptive\", ...(display !== undefined ? { display } : {}) },\n\t\t\t\t\toutput_config: { effort: mapThinkingLevelToEffort(model, options.reasoning) },\n\t\t\t\t}\n\t\t\t: (() => {\n\t\t\t\t\tconst defaultBudgets: Record<ThinkingLevel, number> = {\n\t\t\t\t\t\tminimal: 1024,\n\t\t\t\t\t\tlow: 2048,\n\t\t\t\t\t\tmedium: 8192,\n\t\t\t\t\t\thigh: 16384,\n\t\t\t\t\t\txhigh: 16384, // Claude doesn't support xhigh, clamp to high\n\t\t\t\t\t};\n\n\t\t\t\t\t// Custom budgets override defaults (xhigh not in ThinkingBudgets, use high)\n\t\t\t\t\tconst level = options.reasoning === \"xhigh\" ? \"high\" : options.reasoning;\n\t\t\t\t\tconst budget = options.thinkingBudgets?.[level] ?? defaultBudgets[options.reasoning];\n\n\t\t\t\t\treturn {\n\t\t\t\t\t\tthinking: {\n\t\t\t\t\t\t\ttype: \"enabled\",\n\t\t\t\t\t\t\tbudget_tokens: budget,\n\t\t\t\t\t\t\t...(display !== undefined ? { display } : {}),\n\t\t\t\t\t\t},\n\t\t\t\t\t};\n\t\t\t\t})();\n\n\t\tif (!supportsAdaptiveThinking(model.id, model.name) && (options.interleavedThinking ?? true)) {\n\t\t\tresult.anthropic_beta = [\"interleaved-thinking-2025-05-14\"];\n\t\t}\n\n\t\treturn result;\n\t}\n\n\treturn undefined;\n}\n\nfunction createImageBlock(mimeType: string, data: string) {\n\tlet format: ImageFormat;\n\tswitch (mimeType) {\n\t\tcase \"image/jpeg\":\n\t\tcase \"image/jpg\":\n\t\t\tformat = ImageFormat.JPEG;\n\t\t\tbreak;\n\t\tcase \"image/png\":\n\t\t\tformat = ImageFormat.PNG;\n\t\t\tbreak;\n\t\tcase \"image/gif\":\n\t\t\tformat = ImageFormat.GIF;\n\t\t\tbreak;\n\t\tcase \"image/webp\":\n\t\t\tformat = ImageFormat.WEBP;\n\t\t\tbreak;\n\t\tdefault:\n\t\t\tthrow new Error(`Unknown image type: ${mimeType}`);\n\t}\n\n\tconst binaryString = atob(data);\n\tconst bytes = new Uint8Array(binaryString.length);\n\tfor (let i = 0; i < binaryString.length; i++) {\n\t\tbytes[i] = binaryString.charCodeAt(i);\n\t}\n\n\treturn { source: { bytes }, format };\n}\n"]}

package/dist/providers/amazon-bedrock.js

@@ -1,4 +1,4 @@
-import { BedrockRuntimeClient, StopReason as BedrockStopReason, CachePointType, CacheTTL, ConversationRole, ConverseStreamCommand, ImageFormat, ToolResultStatus, } from "@aws-sdk/client-bedrock-runtime";
+import { BedrockRuntimeClient, BedrockRuntimeServiceException, StopReason as BedrockStopReason, CachePointType, CacheTTL, ConversationRole, ConverseStreamCommand, ImageFormat, ToolResultStatus, } from "@aws-sdk/client-bedrock-runtime";
 import { calculateCost } from "../models.js";
 import { AssistantMessageEventStream } from "../utils/event-stream.js";
 import { parseStreamingJson } from "../utils/json-parse.js";
@@ -29,16 +29,31 @@ export const streamBedrock = (model, context, options = {}) => {
         const config = {
             profile: options.profile,
         };
+        const configuredRegion = getConfiguredBedrockRegion(options);
+        const hasConfiguredProfile = hasConfiguredBedrockProfile();
+        const endpointRegion = getStandardBedrockEndpointRegion(model.baseUrl);
+        const useExplicitEndpoint = shouldUseExplicitBedrockEndpoint(model.baseUrl, configuredRegion, hasConfiguredProfile);
+        // Only pin standard AWS Bedrock runtime endpoints when no region/profile is configured.
+        // This preserves custom endpoints (VPC/proxy) from #3402 without forcing built-in
+        // catalog defaults such as us-east-1 to override AWS_REGION/AWS_PROFILE.
+        if (useExplicitEndpoint) {
+            config.endpoint = model.baseUrl;
+        }
+        // Resolve bearer token for Bedrock API key auth.
+        const bearerToken = options.bearerToken || process.env.AWS_BEARER_TOKEN_BEDROCK || undefined;
+        const useBearerToken = bearerToken !== undefined && process.env.AWS_BEDROCK_SKIP_AUTH !== "1";
         // in Node.js/Bun environment only
         if (typeof process !== "undefined" && (process.versions?.node || process.versions?.bun)) {
             // Region resolution: explicit option > env vars > SDK default chain.
             // When AWS_PROFILE is set, we leave region undefined so the SDK can
             // resovle it from aws profile configs. Otherwise fall back to us-east-1.
-            const explicitRegion = options.region || process.env.AWS_REGION || process.env.AWS_DEFAULT_REGION;
-            if (explicitRegion) {
-                config.region = explicitRegion;
+            if (configuredRegion) {
+                config.region = configuredRegion;
+            }
+            else if (endpointRegion && useExplicitEndpoint) {
+                config.region = endpointRegion;
             }
-            else if (!process.env.AWS_PROFILE) {
+            else if (!hasConfiguredProfile) {
                 config.region = "us-east-1";
             }
             // Support proxies that don't need authentication
@@ -74,22 +89,41 @@ export const streamBedrock = (model, context, options = {}) => {
         else {
             // Non-Node environment (browser): fall back to us-east-1 since
             // there's no config file resolution available.
-            config.region = options.region || "us-east-1";
+            config.region =
+                configuredRegion || (endpointRegion && useExplicitEndpoint ? endpointRegion : undefined) || "us-east-1";
+        }
+        if (useBearerToken) {
+            config.token = { token: bearerToken };
+            config.authSchemePreference = ["httpBearerAuth"];
         }
         try {
             const client = new BedrockRuntimeClient(config);
             const cacheRetention = resolveCacheRetention(options.cacheRetention);
-            const commandInput = {
+            let commandInput = {
                 modelId: model.id,
                 messages: convertMessages(context, model, cacheRetention),
                 system: buildSystemPrompt(context.systemPrompt, model, cacheRetention),
-                inferenceConfig: { maxTokens: options.maxTokens, temperature: options.temperature },
+                inferenceConfig: {
+                    ...(options.maxTokens !== undefined && { maxTokens: options.maxTokens }),
+                    ...(options.temperature !== undefined && { temperature: options.temperature }),
+                },
                 toolConfig: convertToolConfig(context.tools, options.toolChoice),
                 additionalModelRequestFields: buildAdditionalModelRequestFields(model, options),
+                ...(options.requestMetadata !== undefined && { requestMetadata: options.requestMetadata }),
             };
-            options?.onPayload?.(commandInput);
+            const nextCommandInput = await options?.onPayload?.(commandInput, model);
+            if (nextCommandInput !== undefined) {
+                commandInput = nextCommandInput;
+            }
             const command = new ConverseStreamCommand(commandInput);
             const response = await client.send(command, { abortSignal: options.signal });
+            if (response.$metadata.httpStatusCode !== undefined) {
+                const responseHeaders = {};
+                if (response.$metadata.requestId) {
+                    responseHeaders["x-amzn-requestid"] = response.$metadata.requestId;
+                }
+                await options?.onResponse?.({ status: response.$metadata.httpStatusCode, headers: responseHeaders }, model);
+            }
             for await (const item of response.stream) {
                 if (item.messageStart) {
                     if (item.messageStart.role !== ConversationRole.ASSISTANT) {
@@ -113,19 +147,19 @@ export const streamBedrock = (model, context, options = {}) => {
                     handleMetadata(item.metadata, model, output);
                 }
                 else if (item.internalServerException) {
-                    throw new Error(`Internal server error: ${item.internalServerException.message}`);
+                    throw item.internalServerException;
                 }
                 else if (item.modelStreamErrorException) {
-                    throw new Error(`Model stream error: ${item.modelStreamErrorException.message}`);
+                    throw item.modelStreamErrorException;
                 }
                 else if (item.validationException) {
-                    throw new Error(`Validation error: ${item.validationException.message}`);
+                    throw item.validationException;
                 }
                 else if (item.throttlingException) {
-                    throw new Error(`Throttling error: ${item.throttlingException.message}`);
+                    throw item.throttlingException;
                 }
                 else if (item.serviceUnavailableException) {
-                    throw new Error(`Service unavailable: ${item.serviceUnavailableException.message}`);
+                    throw item.serviceUnavailableException;
                 }
             }
             if (options.signal?.aborted) {
@@ -140,23 +174,52 @@ export const streamBedrock = (model, context, options = {}) => {
         catch (error) {
             for (const block of output.content) {
                 delete block.index;
+                // partialJson is only a streaming scratch buffer; never persist it.
                 delete block.partialJson;
             }
             output.stopReason = options.signal?.aborted ? "aborted" : "error";
-            output.errorMessage = error instanceof Error ? error.message : JSON.stringify(error);
+            output.errorMessage = formatBedrockError(error);
             stream.push({ type: "error", reason: output.stopReason, error: output });
             stream.end();
         }
     })();
     return stream;
 };
+/**
+ * Human-readable prefixes for Bedrock SDK exception names.
+ * The downstream retry logic in agent-session matches patterns like
+ * `server.?error` and `service.?unavailable`, so we preserve the legacy
+ * prefix format rather than using the raw SDK exception name.
+ */
+const BEDROCK_ERROR_PREFIXES = {
+    InternalServerException: "Internal server error",
+    ModelStreamErrorException: "Model stream error",
+    ValidationException: "Validation error",
+    ThrottlingException: "Throttling error",
+    ServiceUnavailableException: "Service unavailable",
+};
+/**
+ * Format a Bedrock error with a human-readable prefix.
+ * AWS SDK exceptions (both from `client.send()` and from stream event items)
+ * extend BedrockRuntimeServiceException. We map the `.name` to a stable
+ * human-readable prefix so downstream consumers (retry logic, context-overflow
+ * detection) can distinguish error categories via simple string matching.
+ */
+function formatBedrockError(error) {
+    const message = error instanceof Error ? error.message : JSON.stringify(error);
+    if (error instanceof BedrockRuntimeServiceException) {
+        const prefix = BEDROCK_ERROR_PREFIXES[error.name] ?? error.name;
+        return `${prefix}: ${message}`;
+    }
+    return message;
+}
 export const streamSimpleBedrock = (model, context, options) => {
     const base = buildBaseOptions(model, options, undefined);
     if (!options?.reasoning) {
         return streamBedrock(model, context, { ...base, reasoning: undefined });
     }
-    if (model.id.includes("anthropic.claude") || model.id.includes("anthropic/claude")) {
-        if (supportsAdaptiveThinking(model.id)) {
+    if (isAnthropicClaudeModel(model)) {
+        if (supportsAdaptiveThinking(model.id, model.name)) {
             return streamBedrock(model, context, {
                 ...base,
                 reasoning: options.reasoning,
@@ -272,21 +335,39 @@ function handleContentBlockStop(event, blocks, output, stream) {
             break;
         case "toolCall":
             block.arguments = parseStreamingJson(block.partialJson);
+            // Finalize in-place and strip the scratch buffer so replay only
+            // carries parsed arguments.
             delete block.partialJson;
             stream.push({ type: "toolcall_end", contentIndex: index, toolCall: block, partial: output });
             break;
     }
 }
 /**
- * Check if the model supports adaptive thinking (Opus 4.6 and Sonnet 4.6).
+ * Check if the model supports adaptive thinking (Opus 4.6+, Sonnet 4.6).
+ * Checks both model ID and model name to support application inference profiles
+ * whose ARNs don't contain the model name.
  */
-function supportsAdaptiveThinking(modelId) {
-    return (modelId.includes("opus-4-6") ||
-        modelId.includes("opus-4.6") ||
-        modelId.includes("sonnet-4-6") ||
-        modelId.includes("sonnet-4.6"));
+function getModelMatchCandidates(modelId, modelName) {
+    const values = modelName ? [modelId, modelName] : [modelId];
+    return values.flatMap((value) => {
+        const lower = value.toLowerCase();
+        return [lower, lower.replace(/[\s_.:]+/g, "-")];
+    });
+}
+function supportsAdaptiveThinking(modelId, modelName) {
+    const candidates = getModelMatchCandidates(modelId, modelName);
+    return candidates.some((s) => s.includes("opus-4-6") || s.includes("opus-4-7") || s.includes("sonnet-4-6"));
+}
+function supportsNativeXhighEffort(model) {
+    const candidates = getModelMatchCandidates(model.id, model.name);
+    return candidates.some((s) => s.includes("opus-4-7"));
 }
-function mapThinkingLevelToEffort(level, modelId) {
+function mapThinkingLevelToEffort(model, level) {
+    if (level === "xhigh" && supportsNativeXhighEffort(model))
+        return "xhigh";
+    const mapped = level ? model.thinkingLevelMap?.[level] : undefined;
+    if (typeof mapped === "string")
+        return mapped;
     switch (level) {
         case "minimal":
         case "low":
@@ -295,8 +376,6 @@ function mapThinkingLevelToEffort(level, modelId) {
             return "medium";
         case "high":
             return "high";
-        case "xhigh":
-            return modelId.includes("opus-4-6") || modelId.includes("opus-4.6") ? "max" : "high";
         default:
             return "high";
     }
@@ -314,23 +393,50 @@ function resolveCacheRetention(cacheRetention) {
     }
     return "short";
 }
+/**
+ * Check if the model is an Anthropic Claude model on Bedrock.
+ * Checks both model ID and model name to support application inference profiles
+ * whose ARNs don't contain the model name.
+ */
+function isAnthropicClaudeModel(model) {
+    const id = model.id.toLowerCase();
+    const name = model.name?.toLowerCase() ?? "";
+    return (id.includes("anthropic.claude") ||
+        id.includes("anthropic/claude") ||
+        name.includes("anthropic.claude") ||
+        name.includes("anthropic/claude") ||
+        name.includes("claude"));
+}
 /**
  * Check if the model supports prompt caching.
  * Supported: Claude 3.5 Haiku, Claude 3.7 Sonnet, Claude 4.x models
+ *
+ * For base models and system-defined inference profiles the model ID / ARN
+ * contains the model name, so we can decide locally.
+ *
+ * For application inference profiles (whose ARNs don't contain the model name),
+ * also checks model.name which is user-controlled via models.json or registerProvider.
+ * As a last resort, set AWS_BEDROCK_FORCE_CACHE=1 to enable cache points.
+ * Amazon Nova models have automatic caching and don't need explicit cache points.
  */
 function supportsPromptCaching(model) {
-    if (model.cost.cacheRead || model.cost.cacheWrite) {
-        return true;
+    const candidates = getModelMatchCandidates(model.id, model.name);
+    const hasClaudeRef = candidates.some((s) => s.includes("claude"));
+    if (!hasClaudeRef) {
+        // Application inference profiles don't contain the model name in the ARN.
+        // Allow users to force cache points via environment variable.
+        if (typeof process !== "undefined" && process.env.AWS_BEDROCK_FORCE_CACHE === "1")
+            return true;
+        return false;
     }
-    const id = model.id.toLowerCase();
     // Claude 4.x models (opus-4, sonnet-4, haiku-4)
-    if (id.includes("claude") && (id.includes("-4-") || id.includes("-4.")))
+    if (candidates.some((s) => s.includes("-4-")))
         return true;
     // Claude 3.7 Sonnet
-    if (id.includes("claude-3-7-sonnet"))
+    if (candidates.some((s) => s.includes("claude-3-7-sonnet")))
         return true;
     // Claude 3.5 Haiku
-    if (id.includes("claude-3-5-haiku"))
+    if (candidates.some((s) => s.includes("claude-3-5-haiku")))
         return true;
     return false;
 }
@@ -339,10 +445,11 @@ function supportsPromptCaching(model) {
  * Only Anthropic Claude models support the signature field.
  * Other models (OpenAI, Qwen, Minimax, Moonshot, etc.) reject it with:
  * "This model doesn't support the reasoningContent.reasoningText.signature field"
+ *
+ * Checks both model ID and model name to support application inference profiles.
  */
 function supportsThinkingSignature(model) {
-    const id = model.id.toLowerCase();
-    return id.includes("anthropic.claude") || id.includes("anthropic/claude");
+    return isAnthropicClaudeModel(model);
 }
 function buildSystemPrompt(systemPrompt, model, cacheRetention) {
     if (!systemPrompt)
@@ -411,11 +518,22 @@ function convertMessages(context, model, cacheRetention) {
                             // For other models, we omit the signature to avoid errors like:
                             // "This model doesn't support the reasoningContent.reasoningText.signature field"
                             if (supportsThinkingSignature(model)) {
-                                contentBlocks.push({
-                                    reasoningContent: {
-                                        reasoningText: { text: sanitizeSurrogates(c.thinking), signature: c.thinkingSignature },
-                                    },
-                                });
+                                // Signatures arrive after thinking deltas. If a partial or externally
+                                // persisted message lacks a signature, Bedrock rejects the replayed
+                                // reasoning block. Fall back to plain text, matching Anthropic.
+                                if (!c.thinkingSignature || c.thinkingSignature.trim().length === 0) {
+                                    contentBlocks.push({ text: sanitizeSurrogates(c.thinking) });
+                                }
+                                else {
+                                    contentBlocks.push({
+                                        reasoningContent: {
+                                            reasoningText: {
+                                                text: sanitizeSurrogates(c.thinking),
+                                                signature: c.thinkingSignature,
+                                            },
+                                        },
+                                    });
+                                }
                             }
                             else {
                                 contentBlocks.push({
@@ -533,15 +651,58 @@ function mapStopReason(reason) {
             return "error";
     }
 }
+function getConfiguredBedrockRegion(options) {
+    if (typeof process === "undefined") {
+        return options.region;
+    }
+    return options.region || process.env.AWS_REGION || process.env.AWS_DEFAULT_REGION || undefined;
+}
+function hasConfiguredBedrockProfile() {
+    if (typeof process === "undefined") {
+        return false;
+    }
+    return Boolean(process.env.AWS_PROFILE);
+}
+function getStandardBedrockEndpointRegion(baseUrl) {
+    if (!baseUrl) {
+        return undefined;
+    }
+    try {
+        const { hostname } = new URL(baseUrl);
+        const match = hostname.toLowerCase().match(/^bedrock-runtime(?:-fips)?\.([a-z0-9-]+)\.amazonaws\.com(?:\.cn)?$/);
+        return match?.[1];
+    }
+    catch {
+        return undefined;
+    }
+}
+function shouldUseExplicitBedrockEndpoint(baseUrl, configuredRegion, hasConfiguredProfile) {
+    const endpointRegion = getStandardBedrockEndpointRegion(baseUrl);
+    if (!endpointRegion) {
+        return true;
+    }
+    return !configuredRegion && !hasConfiguredProfile;
+}
+function isGovCloudBedrockTarget(model, options) {
+    const region = getConfiguredBedrockRegion(options);
+    if (region?.toLowerCase().startsWith("us-gov-")) {
+        return true;
+    }
+    const modelId = model.id.toLowerCase();
+    return modelId.startsWith("us-gov.") || modelId.startsWith("arn:aws-us-gov:");
+}
 function buildAdditionalModelRequestFields(model, options) {
     if (!options.reasoning || !model.reasoning) {
         return undefined;
     }
-    if (model.id.includes("anthropic.claude") || model.id.includes("anthropic/claude")) {
-        const result = supportsAdaptiveThinking(model.id)
+    if (isAnthropicClaudeModel(model)) {
+        // GovCloud Bedrock currently rejects the Claude thinking.display field.
+        // Omit it there until the GovCloud Converse schema catches up.
+        const display = isGovCloudBedrockTarget(model, options) ? undefined : (options.thinkingDisplay ?? "summarized");
+        const result = supportsAdaptiveThinking(model.id, model.name)
             ? {
-                thinking: { type: "adaptive" },
-                output_config: { effort: mapThinkingLevelToEffort(options.reasoning, model.id) },
+                thinking: { type: "adaptive", ...(display !== undefined ? { display } : {}) },
+                output_config: { effort: mapThinkingLevelToEffort(model, options.reasoning) },
             }
             : (() => {
                 const defaultBudgets = {
@@ -558,10 +719,11 @@ function buildAdditionalModelRequestFields(model, options) {
                     thinking: {
                         type: "enabled",
                         budget_tokens: budget,
+                        ...(display !== undefined ? { display } : {}),
                     },
                 };
             })();
-        if (!supportsAdaptiveThinking(model.id) && (options.interleavedThinking ?? true)) {
+        if (!supportsAdaptiveThinking(model.id, model.name) && (options.interleavedThinking ?? true)) {
             result.anthropic_beta = ["interleaved-thinking-2025-05-14"];
         }
         return result;