@copilotkit/aimock 1.37.3 → 1.38.0

package/dist/ws-responses.cjs.map

@@ -1 +1 @@
-{"version":3,"file":"ws-responses.cjs","names":["responsesToCompletionRequest","DEFAULT_TEST_ID","matchFixtureDiagnostic","resolveStrictMode","strictNoMatchMessage","strictNoMatchLogLine","flattenHeaders","strictOverrideField","resolveResponse","isErrorResponse","isContentWithToolCallsResponse","buildContentWithToolCallsStreamEvents","resolveReasoningForModel","extractOverrides","createInterruptionSignal","isTextResponse","buildTextStreamEvents","isToolCallResponse","buildToolCallStreamEvents","calculateDelay","delay"],"sources":["../src/ws-responses.ts"],"sourcesContent":["/**\n * WebSocket handler for OpenAI Responses API.\n *\n * Accepts `{ type: \"response.create\", model: \"...\", input: [...] }` messages over\n * WebSocket and sends back the same Responses API SSE events as the HTTP\n * handler, but as individual WebSocket text frames.\n */\n\nimport type { ChatCompletionRequest, Fixture } from \"./types.js\";\nimport { matchFixtureDiagnostic } from \"./router.js\";\nimport {\n  responsesToCompletionRequest,\n  buildTextStreamEvents,\n  buildToolCallStreamEvents,\n  buildContentWithToolCallsStreamEvents,\n  type ResponsesSSEEvent,\n} from \"./responses.js\";\nimport {\n  isTextResponse,\n  isToolCallResponse,\n  isContentWithToolCallsResponse,\n  isErrorResponse,\n  extractOverrides,\n  resolveResponse,\n  resolveStrictMode,\n  resolveReasoningForModel,\n  strictOverrideField,\n  flattenHeaders,\n  strictNoMatchMessage,\n  strictNoMatchLogLine,\n} from \"./helpers.js\";\nimport { createInterruptionSignal } from \"./interruption.js\";\nimport { delay, calculateDelay } from \"./sse-writer.js\";\nimport { DEFAULT_TEST_ID, type Journal } from \"./journal.js\";\nimport type { Logger } from \"./logger.js\";\nimport type { WebSocketConnection } from \"./ws-framing.js\";\n\ninterface ResponseCreateMessage {\n  type: \"response.create\";\n  model?: string;\n  input?: unknown[];\n  instructions?: string;\n  tools?: unknown[];\n  tool_choice?: string | object;\n  stream?: boolean;\n  temperature?: number;\n  max_output_tokens?: number;\n  [key: string]: unknown;\n}\n\nfunction isResponseCreateMessage(msg: unknown): msg is ResponseCreateMessage {\n  return (\n    typeof msg === \"object\" &&\n    msg !== null &&\n    (msg as ResponseCreateMessage).type === \"response.create\"\n  );\n}\n\nfunction buildErrorEvent(\n  message: string,\n  type = \"invalid_request_error\",\n  code?: string,\n): ResponsesSSEEvent {\n  return {\n    type: \"error\",\n    error: { message, type, code },\n  };\n}\n\nexport function handleWebSocketResponses(\n  ws: WebSocketConnection,\n  fixtures: Fixture[],\n  journal: Journal,\n  defaults: {\n    latency: number;\n    chunkSize: number;\n    replaySpeed?: number;\n    model: string;\n    logger: Logger;\n    strict?: boolean;\n    requestTransform?: (req: ChatCompletionRequest) => ChatCompletionRequest;\n    testId?: string;\n    upgradeHeaders?: import(\"node:http\").IncomingHttpHeaders;\n  },\n): void {\n  const { logger } = defaults;\n  // Serialize message processing to prevent event interleaving\n  let pending = Promise.resolve();\n  ws.on(\"message\", (raw: string) => {\n    pending = pending.then(() =>\n      processMessage(raw, ws, fixtures, journal, defaults).catch((err: unknown) => {\n        const msg = err instanceof Error ? err.message : \"Internal error\";\n        logger.error(`WebSocket responses error: ${msg}`);\n        try {\n          ws.send(JSON.stringify(buildErrorEvent(msg, \"server_error\")));\n        } catch (sendErr) {\n          defaults.logger.debug(\n            `Failed to send error to client: ${sendErr instanceof Error ? sendErr.message : \"unknown\"}`,\n          );\n        }\n      }),\n    );\n  });\n}\n\nasync function processMessage(\n  raw: string,\n  ws: WebSocketConnection,\n  fixtures: Fixture[],\n  journal: Journal,\n  defaults: {\n    latency: number;\n    chunkSize: number;\n    replaySpeed?: number;\n    model: string;\n    logger: Logger;\n    strict?: boolean;\n    requestTransform?: (req: ChatCompletionRequest) => ChatCompletionRequest;\n    testId?: string;\n    upgradeHeaders?: import(\"node:http\").IncomingHttpHeaders;\n  },\n): Promise<void> {\n  let parsed: unknown;\n  try {\n    parsed = JSON.parse(raw);\n  } catch (parseErr) {\n    const detail = parseErr instanceof Error ? parseErr.message : \"unknown\";\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(`Malformed JSON: ${detail}`, \"invalid_request_error\", \"invalid_json\"),\n      ),\n    );\n    return;\n  }\n\n  if (!isResponseCreateMessage(parsed)) {\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(\n          'Expected message type \"response.create\"',\n          \"invalid_request_error\",\n          \"invalid_message_type\",\n        ),\n      ),\n    );\n    return;\n  }\n\n  const responsesReq = {\n    model: parsed.model ?? defaults.model,\n    input: (parsed.input ?? []) as {\n      role?: string;\n      type?: string;\n      content?: string | { type: string; text?: string }[];\n      call_id?: string;\n      name?: string;\n      arguments?: string;\n      output?: string;\n      id?: string;\n    }[],\n    instructions: parsed.instructions,\n    tools: parsed.tools as\n      | {\n          type: \"function\";\n          name: string;\n          description?: string;\n          parameters?: object;\n          strict?: boolean;\n        }[]\n      | undefined,\n    tool_choice: parsed.tool_choice,\n    stream: parsed.stream,\n    temperature: parsed.temperature,\n    max_output_tokens: parsed.max_output_tokens,\n  };\n\n  const completionReq = responsesToCompletionRequest(responsesReq);\n  completionReq._endpointType = \"chat\";\n  const contextHeader = defaults.upgradeHeaders?.[\"x-aimock-context\"];\n  completionReq._context =\n    typeof contextHeader === \"string\"\n      ? contextHeader\n      : Array.isArray(contextHeader) && contextHeader.length > 0\n        ? contextHeader[0]\n        : undefined;\n  const testId = defaults.testId ?? DEFAULT_TEST_ID;\n  const { fixture, skippedBySequenceOrTurn } = matchFixtureDiagnostic(\n    fixtures,\n    completionReq,\n    journal.getFixtureMatchCountsForTest(testId),\n    defaults.requestTransform,\n  );\n\n  if (fixture) {\n    journal.incrementFixtureMatchCount(fixture, fixtures, testId);\n  }\n\n  if (!fixture) {\n    if (resolveStrictMode(defaults.strict, defaults.upgradeHeaders)) {\n      const strictMessage = strictNoMatchMessage(skippedBySequenceOrTurn);\n      defaults.logger.error(strictNoMatchLogLine(\"WS\", \"/v1/responses\", skippedBySequenceOrTurn));\n      journal.add({\n        method: \"WS\",\n        path: \"/v1/responses\",\n        headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n        body: completionReq,\n        response: {\n          status: 503,\n          fixture: null,\n          ...strictOverrideField(defaults.strict, defaults.upgradeHeaders),\n        },\n      });\n      ws.close(1008, strictMessage);\n      return;\n    }\n    journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: {\n        status: 404,\n        fixture: null,\n        ...strictOverrideField(defaults.strict, defaults.upgradeHeaders),\n      },\n    });\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(\"No fixture matched\", \"invalid_request_error\", \"no_fixture_match\"),\n      ),\n    );\n    return;\n  }\n\n  const response = await resolveResponse(fixture, completionReq);\n  const latency = fixture.latency ?? defaults.latency;\n  const chunkSize = Math.max(1, fixture.chunkSize ?? defaults.chunkSize);\n\n  // The WS path has no per-request `req.headers`; strict is resolved from the\n  // connection's upgrade headers (see the `!fixture` branch above). Used below\n  // to gate the synthesized reasoning channel on the requested model's capability.\n  const effectiveStrict = resolveStrictMode(defaults.strict, defaults.upgradeHeaders);\n\n  // Error response\n  if (isErrorResponse(response)) {\n    const status = response.status ?? 500;\n    journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status, fixture },\n    });\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(response.error.message, response.error.type, response.error.code),\n      ),\n    );\n    return;\n  }\n\n  // Content + tool calls response (must be checked before isTextResponse / isToolCallResponse)\n  if (isContentWithToolCallsResponse(response)) {\n    const journalEntry = journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status: 200, fixture },\n    });\n\n    const events = buildContentWithToolCallsStreamEvents(\n      response.content ?? \"\",\n      response.toolCalls ?? [],\n      completionReq.model,\n      chunkSize,\n      resolveReasoningForModel(\n        response.reasoning,\n        completionReq.model,\n        effectiveStrict,\n        defaults.logger,\n      ),\n      response.webSearches,\n      extractOverrides(response),\n      response.blocks,\n    );\n\n    const interruption = createInterruptionSignal(fixture);\n    const completed = await sendEvents(\n      ws,\n      events,\n      latency,\n      interruption?.signal,\n      interruption?.tick,\n      fixture.recordedTimings,\n      fixture.replaySpeed ?? defaults.replaySpeed,\n    );\n    if (!completed) {\n      ws.destroy();\n      journalEntry.response.interrupted = true;\n      journalEntry.response.interruptReason = interruption?.reason();\n    }\n    interruption?.cleanup();\n    return;\n  }\n\n  // Text response\n  if (isTextResponse(response)) {\n    const journalEntry = journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status: 200, fixture },\n    });\n\n    const events = buildTextStreamEvents(\n      response.content,\n      completionReq.model,\n      chunkSize,\n      resolveReasoningForModel(\n        response.reasoning,\n        completionReq.model,\n        effectiveStrict,\n        defaults.logger,\n      ),\n      response.webSearches,\n      extractOverrides(response),\n    );\n    const interruption = createInterruptionSignal(fixture);\n    const completed = await sendEvents(\n      ws,\n      events,\n      latency,\n      interruption?.signal,\n      interruption?.tick,\n      fixture.recordedTimings,\n      fixture.replaySpeed ?? defaults.replaySpeed,\n    );\n    if (!completed) {\n      ws.destroy();\n      journalEntry.response.interrupted = true;\n      journalEntry.response.interruptReason = interruption?.reason();\n    }\n    interruption?.cleanup();\n    return;\n  }\n\n  // Tool call response\n  if (isToolCallResponse(response)) {\n    const journalEntry = journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status: 200, fixture },\n    });\n    const events = buildToolCallStreamEvents(\n      response.toolCalls,\n      completionReq.model,\n      chunkSize,\n      // Gate the synthesized reasoning channel on the requested model's\n      // capability, matching the WS text / content+tool branches and the HTTP\n      // tool-only path so reasoning emission is transport-independent.\n      resolveReasoningForModel(\n        response.reasoning,\n        completionReq.model,\n        effectiveStrict,\n        defaults.logger,\n      ),\n      response.webSearches,\n      extractOverrides(response),\n    );\n    const interruption = createInterruptionSignal(fixture);\n    const completed = await sendEvents(\n      ws,\n      events,\n      latency,\n      interruption?.signal,\n      interruption?.tick,\n      fixture.recordedTimings,\n      fixture.replaySpeed ?? defaults.replaySpeed,\n    );\n    if (!completed) {\n      ws.destroy();\n      journalEntry.response.interrupted = true;\n      journalEntry.response.interruptReason = interruption?.reason();\n    }\n    interruption?.cleanup();\n    return;\n  }\n\n  // Unknown response type\n  journal.add({\n    method: \"WS\",\n    path: \"/v1/responses\",\n    headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n    body: completionReq,\n    response: { status: 500, fixture },\n  });\n  ws.send(\n    JSON.stringify(\n      buildErrorEvent(\"Fixture response did not match any known type\", \"server_error\"),\n    ),\n  );\n}\n\nasync function sendEvents(\n  ws: WebSocketConnection,\n  events: ResponsesSSEEvent[],\n  latency: number,\n  signal?: AbortSignal,\n  onChunkSent?: () => void,\n  recordedTimings?: import(\"./types.js\").RecordedTimings,\n  replaySpeed?: number,\n): Promise<boolean> {\n  let eventIndex = 0;\n  for (const event of events) {\n    if (ws.isClosed) return false;\n    const chunkDelay = calculateDelay(eventIndex, undefined, latency, recordedTimings, replaySpeed);\n    if (chunkDelay > 0) await delay(chunkDelay, signal);\n    if (signal?.aborted) return false;\n    if (ws.isClosed) return false;\n    ws.send(JSON.stringify(event));\n    eventIndex++;\n    onChunkSent?.();\n    if (signal?.aborted) return false;\n  }\n  return true;\n}\n"],"mappings":";;;;;;;;;AAkDA,SAAS,wBAAwB,KAA4C;AAC3E,QACE,OAAO,QAAQ,YACf,QAAQ,QACP,IAA8B,SAAS;;AAI5C,SAAS,gBACP,SACA,OAAO,yBACP,MACmB;AACnB,QAAO;EACL,MAAM;EACN,OAAO;GAAE;GAAS;GAAM;GAAM;EAC/B;;AAGH,SAAgB,yBACd,IACA,UACA,SACA,UAWM;CACN,MAAM,EAAE,WAAW;CAEnB,IAAI,UAAU,QAAQ,SAAS;AAC/B,IAAG,GAAG,YAAY,QAAgB;AAChC,YAAU,QAAQ,WAChB,eAAe,KAAK,IAAI,UAAU,SAAS,SAAS,CAAC,OAAO,QAAiB;GAC3E,MAAM,MAAM,eAAe,QAAQ,IAAI,UAAU;AACjD,UAAO,MAAM,8BAA8B,MAAM;AACjD,OAAI;AACF,OAAG,KAAK,KAAK,UAAU,gBAAgB,KAAK,eAAe,CAAC,CAAC;YACtD,SAAS;AAChB,aAAS,OAAO,MACd,mCAAmC,mBAAmB,QAAQ,QAAQ,UAAU,YACjF;;IAEH,CACH;GACD;;AAGJ,eAAe,eACb,KACA,IACA,UACA,SACA,UAWe;CACf,IAAI;AACJ,KAAI;AACF,WAAS,KAAK,MAAM,IAAI;UACjB,UAAU;EACjB,MAAM,SAAS,oBAAoB,QAAQ,SAAS,UAAU;AAC9D,KAAG,KACD,KAAK,UACH,gBAAgB,mBAAmB,UAAU,yBAAyB,eAAe,CACtF,CACF;AACD;;AAGF,KAAI,CAAC,wBAAwB,OAAO,EAAE;AACpC,KAAG,KACD,KAAK,UACH,gBACE,6CACA,yBACA,uBACD,CACF,CACF;AACD;;CA+BF,MAAM,gBAAgBA,+CA5BD;EACnB,OAAO,OAAO,SAAS,SAAS;EAChC,OAAQ,OAAO,SAAS,EAAE;EAU1B,cAAc,OAAO;EACrB,OAAO,OAAO;EASd,aAAa,OAAO;EACpB,QAAQ,OAAO;EACf,aAAa,OAAO;EACpB,mBAAmB,OAAO;EAC3B,CAE+D;AAChE,eAAc,gBAAgB;CAC9B,MAAM,gBAAgB,SAAS,iBAAiB;AAChD,eAAc,WACZ,OAAO,kBAAkB,WACrB,gBACA,MAAM,QAAQ,cAAc,IAAI,cAAc,SAAS,IACrD,cAAc,KACd;CACR,MAAM,SAAS,SAAS,UAAUC;CAClC,MAAM,EAAE,SAAS,4BAA4BC,sCAC3C,UACA,eACA,QAAQ,6BAA6B,OAAO,EAC5C,SAAS,iBACV;AAED,KAAI,QACF,SAAQ,2BAA2B,SAAS,UAAU,OAAO;AAG/D,KAAI,CAAC,SAAS;AACZ,MAAIC,kCAAkB,SAAS,QAAQ,SAAS,eAAe,EAAE;GAC/D,MAAM,gBAAgBC,qCAAqB,wBAAwB;AACnE,YAAS,OAAO,MAAMC,qCAAqB,MAAM,iBAAiB,wBAAwB,CAAC;AAC3F,WAAQ,IAAI;IACV,QAAQ;IACR,MAAM;IACN,SAASC,+BAAe,SAAS,kBAAkB,EAAE,CAAC;IACtD,MAAM;IACN,UAAU;KACR,QAAQ;KACR,SAAS;KACT,GAAGC,oCAAoB,SAAS,QAAQ,SAAS,eAAe;KACjE;IACF,CAAC;AACF,MAAG,MAAM,MAAM,cAAc;AAC7B;;AAEF,UAAQ,IAAI;GACV,QAAQ;GACR,MAAM;GACN,SAASD,+BAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IACR,QAAQ;IACR,SAAS;IACT,GAAGC,oCAAoB,SAAS,QAAQ,SAAS,eAAe;IACjE;GACF,CAAC;AACF,KAAG,KACD,KAAK,UACH,gBAAgB,sBAAsB,yBAAyB,mBAAmB,CACnF,CACF;AACD;;CAGF,MAAM,WAAW,MAAMC,gCAAgB,SAAS,cAAc;CAC9D,MAAM,UAAU,QAAQ,WAAW,SAAS;CAC5C,MAAM,YAAY,KAAK,IAAI,GAAG,QAAQ,aAAa,SAAS,UAAU;CAKtE,MAAM,kBAAkBL,kCAAkB,SAAS,QAAQ,SAAS,eAAe;AAGnF,KAAIM,gCAAgB,SAAS,EAAE;EAC7B,MAAM,SAAS,SAAS,UAAU;AAClC,UAAQ,IAAI;GACV,QAAQ;GACR,MAAM;GACN,SAASH,+BAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE;IAAQ;IAAS;GAC9B,CAAC;AACF,KAAG,KACD,KAAK,UACH,gBAAgB,SAAS,MAAM,SAAS,SAAS,MAAM,MAAM,SAAS,MAAM,KAAK,CAClF,CACF;AACD;;AAIF,KAAII,+CAA+B,SAAS,EAAE;EAC5C,MAAM,eAAe,QAAQ,IAAI;GAC/B,QAAQ;GACR,MAAM;GACN,SAASJ,+BAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE,QAAQ;IAAK;IAAS;GACnC,CAAC;EAEF,MAAM,SAASK,wDACb,SAAS,WAAW,IACpB,SAAS,aAAa,EAAE,EACxB,cAAc,OACd,WACAC,yCACE,SAAS,WACT,cAAc,OACd,iBACA,SAAS,OACV,EACD,SAAS,aACTC,iCAAiB,SAAS,EAC1B,SAAS,OACV;EAED,MAAM,eAAeC,8CAAyB,QAAQ;AAUtD,MAAI,CATc,MAAM,WACtB,IACA,QACA,SACA,cAAc,QACd,cAAc,MACd,QAAQ,iBACR,QAAQ,eAAe,SAAS,YACjC,EACe;AACd,MAAG,SAAS;AACZ,gBAAa,SAAS,cAAc;AACpC,gBAAa,SAAS,kBAAkB,cAAc,QAAQ;;AAEhE,gBAAc,SAAS;AACvB;;AAIF,KAAIC,+BAAe,SAAS,EAAE;EAC5B,MAAM,eAAe,QAAQ,IAAI;GAC/B,QAAQ;GACR,MAAM;GACN,SAAST,+BAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE,QAAQ;IAAK;IAAS;GACnC,CAAC;EAEF,MAAM,SAASU,wCACb,SAAS,SACT,cAAc,OACd,WACAJ,yCACE,SAAS,WACT,cAAc,OACd,iBACA,SAAS,OACV,EACD,SAAS,aACTC,iCAAiB,SAAS,CAC3B;EACD,MAAM,eAAeC,8CAAyB,QAAQ;AAUtD,MAAI,CATc,MAAM,WACtB,IACA,QACA,SACA,cAAc,QACd,cAAc,MACd,QAAQ,iBACR,QAAQ,eAAe,SAAS,YACjC,EACe;AACd,MAAG,SAAS;AACZ,gBAAa,SAAS,cAAc;AACpC,gBAAa,SAAS,kBAAkB,cAAc,QAAQ;;AAEhE,gBAAc,SAAS;AACvB;;AAIF,KAAIG,mCAAmB,SAAS,EAAE;EAChC,MAAM,eAAe,QAAQ,IAAI;GAC/B,QAAQ;GACR,MAAM;GACN,SAASX,+BAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE,QAAQ;IAAK;IAAS;GACnC,CAAC;EACF,MAAM,SAASY,4CACb,SAAS,WACT,cAAc,OACd,WAIAN,yCACE,SAAS,WACT,cAAc,OACd,iBACA,SAAS,OACV,EACD,SAAS,aACTC,iCAAiB,SAAS,CAC3B;EACD,MAAM,eAAeC,8CAAyB,QAAQ;AAUtD,MAAI,CATc,MAAM,WACtB,IACA,QACA,SACA,cAAc,QACd,cAAc,MACd,QAAQ,iBACR,QAAQ,eAAe,SAAS,YACjC,EACe;AACd,MAAG,SAAS;AACZ,gBAAa,SAAS,cAAc;AACpC,gBAAa,SAAS,kBAAkB,cAAc,QAAQ;;AAEhE,gBAAc,SAAS;AACvB;;AAIF,SAAQ,IAAI;EACV,QAAQ;EACR,MAAM;EACN,SAASR,+BAAe,SAAS,kBAAkB,EAAE,CAAC;EACtD,MAAM;EACN,UAAU;GAAE,QAAQ;GAAK;GAAS;EACnC,CAAC;AACF,IAAG,KACD,KAAK,UACH,gBAAgB,iDAAiD,eAAe,CACjF,CACF;;AAGH,eAAe,WACb,IACA,QACA,SACA,QACA,aACA,iBACA,aACkB;CAClB,IAAI,aAAa;AACjB,MAAK,MAAM,SAAS,QAAQ;AAC1B,MAAI,GAAG,SAAU,QAAO;EACxB,MAAM,aAAaa,kCAAe,YAAY,QAAW,SAAS,iBAAiB,YAAY;AAC/F,MAAI,aAAa,EAAG,OAAMC,yBAAM,YAAY,OAAO;AACnD,MAAI,QAAQ,QAAS,QAAO;AAC5B,MAAI,GAAG,SAAU,QAAO;AACxB,KAAG,KAAK,KAAK,UAAU,MAAM,CAAC;AAC9B;AACA,iBAAe;AACf,MAAI,QAAQ,QAAS,QAAO;;AAE9B,QAAO"}
+{"version":3,"file":"ws-responses.cjs","names":["requestWantsEncryptedReasoning","requestIncludesEncryptedReasoning","responsesToCompletionRequest","DEFAULT_TEST_ID","matchFixtureDiagnostic","resolveStrictMode","strictNoMatchMessage","strictNoMatchLogLine","flattenHeaders","strictOverrideField","resolveResponse","isErrorResponse","isContentWithToolCallsResponse","buildContentWithToolCallsStreamEvents","resolveReasoningForModel","extractOverrides","createInterruptionSignal","isTextResponse","buildTextStreamEvents","isToolCallResponse","buildToolCallStreamEvents","calculateDelay","delay"],"sources":["../src/ws-responses.ts"],"sourcesContent":["/**\n * WebSocket handler for OpenAI Responses API.\n *\n * Accepts `{ type: \"response.create\", model: \"...\", input: [...] }` messages over\n * WebSocket and sends back the same Responses API SSE events as the HTTP\n * handler, but as individual WebSocket text frames.\n */\n\nimport type { ChatCompletionRequest, Fixture } from \"./types.js\";\nimport { matchFixtureDiagnostic } from \"./router.js\";\nimport {\n  responsesToCompletionRequest,\n  requestIncludesEncryptedReasoning,\n  requestWantsEncryptedReasoning,\n  buildTextStreamEvents,\n  buildToolCallStreamEvents,\n  buildContentWithToolCallsStreamEvents,\n  type ResponsesSSEEvent,\n} from \"./responses.js\";\nimport {\n  isTextResponse,\n  isToolCallResponse,\n  isContentWithToolCallsResponse,\n  isErrorResponse,\n  extractOverrides,\n  resolveResponse,\n  resolveStrictMode,\n  resolveReasoningForModel,\n  strictOverrideField,\n  flattenHeaders,\n  strictNoMatchMessage,\n  strictNoMatchLogLine,\n} from \"./helpers.js\";\nimport { createInterruptionSignal } from \"./interruption.js\";\nimport { delay, calculateDelay } from \"./sse-writer.js\";\nimport { DEFAULT_TEST_ID, type Journal } from \"./journal.js\";\nimport type { Logger } from \"./logger.js\";\nimport type { WebSocketConnection } from \"./ws-framing.js\";\n\ninterface ResponseCreateMessage {\n  type: \"response.create\";\n  model?: string;\n  input?: unknown[];\n  instructions?: string;\n  tools?: unknown[];\n  tool_choice?: string | object;\n  stream?: boolean;\n  temperature?: number;\n  max_output_tokens?: number;\n  [key: string]: unknown;\n}\n\nfunction isResponseCreateMessage(msg: unknown): msg is ResponseCreateMessage {\n  return (\n    typeof msg === \"object\" &&\n    msg !== null &&\n    (msg as ResponseCreateMessage).type === \"response.create\"\n  );\n}\n\nfunction buildErrorEvent(\n  message: string,\n  type = \"invalid_request_error\",\n  code?: string,\n): ResponsesSSEEvent {\n  return {\n    type: \"error\",\n    error: { message, type, code },\n  };\n}\n\nexport function handleWebSocketResponses(\n  ws: WebSocketConnection,\n  fixtures: Fixture[],\n  journal: Journal,\n  defaults: {\n    latency: number;\n    chunkSize: number;\n    replaySpeed?: number;\n    model: string;\n    logger: Logger;\n    strict?: boolean;\n    requestTransform?: (req: ChatCompletionRequest) => ChatCompletionRequest;\n    testId?: string;\n    upgradeHeaders?: import(\"node:http\").IncomingHttpHeaders;\n  },\n): void {\n  const { logger } = defaults;\n  // Serialize message processing to prevent event interleaving\n  let pending = Promise.resolve();\n  ws.on(\"message\", (raw: string) => {\n    pending = pending.then(() =>\n      processMessage(raw, ws, fixtures, journal, defaults).catch((err: unknown) => {\n        const msg = err instanceof Error ? err.message : \"Internal error\";\n        logger.error(`WebSocket responses error: ${msg}`);\n        try {\n          ws.send(JSON.stringify(buildErrorEvent(msg, \"server_error\")));\n        } catch (sendErr) {\n          defaults.logger.debug(\n            `Failed to send error to client: ${sendErr instanceof Error ? sendErr.message : \"unknown\"}`,\n          );\n        }\n      }),\n    );\n  });\n}\n\nasync function processMessage(\n  raw: string,\n  ws: WebSocketConnection,\n  fixtures: Fixture[],\n  journal: Journal,\n  defaults: {\n    latency: number;\n    chunkSize: number;\n    replaySpeed?: number;\n    model: string;\n    logger: Logger;\n    strict?: boolean;\n    requestTransform?: (req: ChatCompletionRequest) => ChatCompletionRequest;\n    testId?: string;\n    upgradeHeaders?: import(\"node:http\").IncomingHttpHeaders;\n  },\n): Promise<void> {\n  let parsed: unknown;\n  try {\n    parsed = JSON.parse(raw);\n  } catch (parseErr) {\n    const detail = parseErr instanceof Error ? parseErr.message : \"unknown\";\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(`Malformed JSON: ${detail}`, \"invalid_request_error\", \"invalid_json\"),\n      ),\n    );\n    return;\n  }\n\n  if (!isResponseCreateMessage(parsed)) {\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(\n          'Expected message type \"response.create\"',\n          \"invalid_request_error\",\n          \"invalid_message_type\",\n        ),\n      ),\n    );\n    return;\n  }\n\n  const responsesReq = {\n    model: parsed.model ?? defaults.model,\n    input: (parsed.input ?? []) as {\n      role?: string;\n      type?: string;\n      content?: string | { type: string; text?: string }[];\n      call_id?: string;\n      name?: string;\n      arguments?: string;\n      output?: string;\n      id?: string;\n    }[],\n    instructions: parsed.instructions,\n    tools: parsed.tools as\n      | {\n          type: \"function\";\n          name: string;\n          description?: string;\n          parameters?: object;\n          strict?: boolean;\n        }[]\n      | undefined,\n    tool_choice: parsed.tool_choice,\n    stream: parsed.stream,\n    temperature: parsed.temperature,\n    max_output_tokens: parsed.max_output_tokens,\n    include: (parsed as { include?: string[] }).include,\n    store: (parsed as { store?: boolean }).store,\n  };\n\n  // Gate encrypted-reasoning emission identically to the HTTP transport so the\n  // agent-framework#7233 stateless-replay path works over WebSocket too.\n  const emitEncryptedReasoning = requestWantsEncryptedReasoning(responsesReq);\n  // Synthesizing a reasoning item the fixture never declared takes the NARROWER\n  // explicit-`include` gate, transport-identically to HTTP — a `store: false`\n  // request with a summary-less fixture still gets no reasoning item.\n  const synthesizeSummarylessReasoning = requestIncludesEncryptedReasoning(responsesReq);\n\n  const completionReq = responsesToCompletionRequest(responsesReq);\n  completionReq._endpointType = \"chat\";\n  const contextHeader = defaults.upgradeHeaders?.[\"x-aimock-context\"];\n  completionReq._context =\n    typeof contextHeader === \"string\"\n      ? contextHeader\n      : Array.isArray(contextHeader) && contextHeader.length > 0\n        ? contextHeader[0]\n        : undefined;\n  const testId = defaults.testId ?? DEFAULT_TEST_ID;\n  const { fixture, skippedBySequenceOrTurn } = matchFixtureDiagnostic(\n    fixtures,\n    completionReq,\n    journal.getFixtureMatchCountsForTest(testId),\n    defaults.requestTransform,\n  );\n\n  if (fixture) {\n    journal.incrementFixtureMatchCount(fixture, fixtures, testId);\n  }\n\n  if (!fixture) {\n    if (resolveStrictMode(defaults.strict, defaults.upgradeHeaders)) {\n      const strictMessage = strictNoMatchMessage(skippedBySequenceOrTurn);\n      defaults.logger.error(strictNoMatchLogLine(\"WS\", \"/v1/responses\", skippedBySequenceOrTurn));\n      journal.add({\n        method: \"WS\",\n        path: \"/v1/responses\",\n        headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n        body: completionReq,\n        response: {\n          status: 503,\n          fixture: null,\n          ...strictOverrideField(defaults.strict, defaults.upgradeHeaders),\n        },\n      });\n      ws.close(1008, strictMessage);\n      return;\n    }\n    journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: {\n        status: 404,\n        fixture: null,\n        ...strictOverrideField(defaults.strict, defaults.upgradeHeaders),\n      },\n    });\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(\"No fixture matched\", \"invalid_request_error\", \"no_fixture_match\"),\n      ),\n    );\n    return;\n  }\n\n  const response = await resolveResponse(fixture, completionReq);\n  const latency = fixture.latency ?? defaults.latency;\n  const chunkSize = Math.max(1, fixture.chunkSize ?? defaults.chunkSize);\n\n  // The WS path has no per-request `req.headers`; strict is resolved from the\n  // connection's upgrade headers (see the `!fixture` branch above). Used below\n  // to gate the synthesized reasoning channel on the requested model's capability.\n  const effectiveStrict = resolveStrictMode(defaults.strict, defaults.upgradeHeaders);\n\n  // Error response\n  if (isErrorResponse(response)) {\n    const status = response.status ?? 500;\n    journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status, fixture },\n    });\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(response.error.message, response.error.type, response.error.code),\n      ),\n    );\n    return;\n  }\n\n  // Content + tool calls response (must be checked before isTextResponse / isToolCallResponse)\n  if (isContentWithToolCallsResponse(response)) {\n    const journalEntry = journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status: 200, fixture },\n    });\n\n    const events = buildContentWithToolCallsStreamEvents(\n      response.content ?? \"\",\n      response.toolCalls ?? [],\n      completionReq.model,\n      chunkSize,\n      resolveReasoningForModel(\n        response.reasoning,\n        completionReq.model,\n        effectiveStrict,\n        defaults.logger,\n      ),\n      response.webSearches,\n      extractOverrides(response),\n      response.blocks,\n      emitEncryptedReasoning,\n      synthesizeSummarylessReasoning,\n    );\n\n    const interruption = createInterruptionSignal(fixture);\n    const completed = await sendEvents(\n      ws,\n      events,\n      latency,\n      interruption?.signal,\n      interruption?.tick,\n      fixture.recordedTimings,\n      fixture.replaySpeed ?? defaults.replaySpeed,\n    );\n    if (!completed) {\n      ws.destroy();\n      journalEntry.response.interrupted = true;\n      journalEntry.response.interruptReason = interruption?.reason();\n    }\n    interruption?.cleanup();\n    return;\n  }\n\n  // Text response\n  if (isTextResponse(response)) {\n    const journalEntry = journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status: 200, fixture },\n    });\n\n    const events = buildTextStreamEvents(\n      response.content,\n      completionReq.model,\n      chunkSize,\n      resolveReasoningForModel(\n        response.reasoning,\n        completionReq.model,\n        effectiveStrict,\n        defaults.logger,\n      ),\n      response.webSearches,\n      extractOverrides(response),\n      emitEncryptedReasoning,\n      synthesizeSummarylessReasoning,\n    );\n    const interruption = createInterruptionSignal(fixture);\n    const completed = await sendEvents(\n      ws,\n      events,\n      latency,\n      interruption?.signal,\n      interruption?.tick,\n      fixture.recordedTimings,\n      fixture.replaySpeed ?? defaults.replaySpeed,\n    );\n    if (!completed) {\n      ws.destroy();\n      journalEntry.response.interrupted = true;\n      journalEntry.response.interruptReason = interruption?.reason();\n    }\n    interruption?.cleanup();\n    return;\n  }\n\n  // Tool call response\n  if (isToolCallResponse(response)) {\n    const journalEntry = journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status: 200, fixture },\n    });\n    const events = buildToolCallStreamEvents(\n      response.toolCalls,\n      completionReq.model,\n      chunkSize,\n      // Gate the synthesized reasoning channel on the requested model's\n      // capability, matching the WS text / content+tool branches and the HTTP\n      // tool-only path so reasoning emission is transport-independent.\n      resolveReasoningForModel(\n        response.reasoning,\n        completionReq.model,\n        effectiveStrict,\n        defaults.logger,\n      ),\n      response.webSearches,\n      extractOverrides(response),\n      emitEncryptedReasoning,\n      synthesizeSummarylessReasoning,\n    );\n    const interruption = createInterruptionSignal(fixture);\n    const completed = await sendEvents(\n      ws,\n      events,\n      latency,\n      interruption?.signal,\n      interruption?.tick,\n      fixture.recordedTimings,\n      fixture.replaySpeed ?? defaults.replaySpeed,\n    );\n    if (!completed) {\n      ws.destroy();\n      journalEntry.response.interrupted = true;\n      journalEntry.response.interruptReason = interruption?.reason();\n    }\n    interruption?.cleanup();\n    return;\n  }\n\n  // Unknown response type\n  journal.add({\n    method: \"WS\",\n    path: \"/v1/responses\",\n    headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n    body: completionReq,\n    response: { status: 500, fixture },\n  });\n  ws.send(\n    JSON.stringify(\n      buildErrorEvent(\"Fixture response did not match any known type\", \"server_error\"),\n    ),\n  );\n}\n\nasync function sendEvents(\n  ws: WebSocketConnection,\n  events: ResponsesSSEEvent[],\n  latency: number,\n  signal?: AbortSignal,\n  onChunkSent?: () => void,\n  recordedTimings?: import(\"./types.js\").RecordedTimings,\n  replaySpeed?: number,\n): Promise<boolean> {\n  let eventIndex = 0;\n  for (const event of events) {\n    if (ws.isClosed) return false;\n    const chunkDelay = calculateDelay(eventIndex, undefined, latency, recordedTimings, replaySpeed);\n    if (chunkDelay > 0) await delay(chunkDelay, signal);\n    if (signal?.aborted) return false;\n    if (ws.isClosed) return false;\n    ws.send(JSON.stringify(event));\n    eventIndex++;\n    onChunkSent?.();\n    if (signal?.aborted) return false;\n  }\n  return true;\n}\n"],"mappings":";;;;;;;;;AAoDA,SAAS,wBAAwB,KAA4C;AAC3E,QACE,OAAO,QAAQ,YACf,QAAQ,QACP,IAA8B,SAAS;;AAI5C,SAAS,gBACP,SACA,OAAO,yBACP,MACmB;AACnB,QAAO;EACL,MAAM;EACN,OAAO;GAAE;GAAS;GAAM;GAAM;EAC/B;;AAGH,SAAgB,yBACd,IACA,UACA,SACA,UAWM;CACN,MAAM,EAAE,WAAW;CAEnB,IAAI,UAAU,QAAQ,SAAS;AAC/B,IAAG,GAAG,YAAY,QAAgB;AAChC,YAAU,QAAQ,WAChB,eAAe,KAAK,IAAI,UAAU,SAAS,SAAS,CAAC,OAAO,QAAiB;GAC3E,MAAM,MAAM,eAAe,QAAQ,IAAI,UAAU;AACjD,UAAO,MAAM,8BAA8B,MAAM;AACjD,OAAI;AACF,OAAG,KAAK,KAAK,UAAU,gBAAgB,KAAK,eAAe,CAAC,CAAC;YACtD,SAAS;AAChB,aAAS,OAAO,MACd,mCAAmC,mBAAmB,QAAQ,QAAQ,UAAU,YACjF;;IAEH,CACH;GACD;;AAGJ,eAAe,eACb,KACA,IACA,UACA,SACA,UAWe;CACf,IAAI;AACJ,KAAI;AACF,WAAS,KAAK,MAAM,IAAI;UACjB,UAAU;EACjB,MAAM,SAAS,oBAAoB,QAAQ,SAAS,UAAU;AAC9D,KAAG,KACD,KAAK,UACH,gBAAgB,mBAAmB,UAAU,yBAAyB,eAAe,CACtF,CACF;AACD;;AAGF,KAAI,CAAC,wBAAwB,OAAO,EAAE;AACpC,KAAG,KACD,KAAK,UACH,gBACE,6CACA,yBACA,uBACD,CACF,CACF;AACD;;CAGF,MAAM,eAAe;EACnB,OAAO,OAAO,SAAS,SAAS;EAChC,OAAQ,OAAO,SAAS,EAAE;EAU1B,cAAc,OAAO;EACrB,OAAO,OAAO;EASd,aAAa,OAAO;EACpB,QAAQ,OAAO;EACf,aAAa,OAAO;EACpB,mBAAmB,OAAO;EAC1B,SAAU,OAAkC;EAC5C,OAAQ,OAA+B;EACxC;CAID,MAAM,yBAAyBA,iDAA+B,aAAa;CAI3E,MAAM,iCAAiCC,oDAAkC,aAAa;CAEtF,MAAM,gBAAgBC,+CAA6B,aAAa;AAChE,eAAc,gBAAgB;CAC9B,MAAM,gBAAgB,SAAS,iBAAiB;AAChD,eAAc,WACZ,OAAO,kBAAkB,WACrB,gBACA,MAAM,QAAQ,cAAc,IAAI,cAAc,SAAS,IACrD,cAAc,KACd;CACR,MAAM,SAAS,SAAS,UAAUC;CAClC,MAAM,EAAE,SAAS,4BAA4BC,sCAC3C,UACA,eACA,QAAQ,6BAA6B,OAAO,EAC5C,SAAS,iBACV;AAED,KAAI,QACF,SAAQ,2BAA2B,SAAS,UAAU,OAAO;AAG/D,KAAI,CAAC,SAAS;AACZ,MAAIC,kCAAkB,SAAS,QAAQ,SAAS,eAAe,EAAE;GAC/D,MAAM,gBAAgBC,qCAAqB,wBAAwB;AACnE,YAAS,OAAO,MAAMC,qCAAqB,MAAM,iBAAiB,wBAAwB,CAAC;AAC3F,WAAQ,IAAI;IACV,QAAQ;IACR,MAAM;IACN,SAASC,+BAAe,SAAS,kBAAkB,EAAE,CAAC;IACtD,MAAM;IACN,UAAU;KACR,QAAQ;KACR,SAAS;KACT,GAAGC,oCAAoB,SAAS,QAAQ,SAAS,eAAe;KACjE;IACF,CAAC;AACF,MAAG,MAAM,MAAM,cAAc;AAC7B;;AAEF,UAAQ,IAAI;GACV,QAAQ;GACR,MAAM;GACN,SAASD,+BAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IACR,QAAQ;IACR,SAAS;IACT,GAAGC,oCAAoB,SAAS,QAAQ,SAAS,eAAe;IACjE;GACF,CAAC;AACF,KAAG,KACD,KAAK,UACH,gBAAgB,sBAAsB,yBAAyB,mBAAmB,CACnF,CACF;AACD;;CAGF,MAAM,WAAW,MAAMC,gCAAgB,SAAS,cAAc;CAC9D,MAAM,UAAU,QAAQ,WAAW,SAAS;CAC5C,MAAM,YAAY,KAAK,IAAI,GAAG,QAAQ,aAAa,SAAS,UAAU;CAKtE,MAAM,kBAAkBL,kCAAkB,SAAS,QAAQ,SAAS,eAAe;AAGnF,KAAIM,gCAAgB,SAAS,EAAE;EAC7B,MAAM,SAAS,SAAS,UAAU;AAClC,UAAQ,IAAI;GACV,QAAQ;GACR,MAAM;GACN,SAASH,+BAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE;IAAQ;IAAS;GAC9B,CAAC;AACF,KAAG,KACD,KAAK,UACH,gBAAgB,SAAS,MAAM,SAAS,SAAS,MAAM,MAAM,SAAS,MAAM,KAAK,CAClF,CACF;AACD;;AAIF,KAAII,+CAA+B,SAAS,EAAE;EAC5C,MAAM,eAAe,QAAQ,IAAI;GAC/B,QAAQ;GACR,MAAM;GACN,SAASJ,+BAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE,QAAQ;IAAK;IAAS;GACnC,CAAC;EAEF,MAAM,SAASK,wDACb,SAAS,WAAW,IACpB,SAAS,aAAa,EAAE,EACxB,cAAc,OACd,WACAC,yCACE,SAAS,WACT,cAAc,OACd,iBACA,SAAS,OACV,EACD,SAAS,aACTC,iCAAiB,SAAS,EAC1B,SAAS,QACT,wBACA,+BACD;EAED,MAAM,eAAeC,8CAAyB,QAAQ;AAUtD,MAAI,CATc,MAAM,WACtB,IACA,QACA,SACA,cAAc,QACd,cAAc,MACd,QAAQ,iBACR,QAAQ,eAAe,SAAS,YACjC,EACe;AACd,MAAG,SAAS;AACZ,gBAAa,SAAS,cAAc;AACpC,gBAAa,SAAS,kBAAkB,cAAc,QAAQ;;AAEhE,gBAAc,SAAS;AACvB;;AAIF,KAAIC,+BAAe,SAAS,EAAE;EAC5B,MAAM,eAAe,QAAQ,IAAI;GAC/B,QAAQ;GACR,MAAM;GACN,SAAST,+BAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE,QAAQ;IAAK;IAAS;GACnC,CAAC;EAEF,MAAM,SAASU,wCACb,SAAS,SACT,cAAc,OACd,WACAJ,yCACE,SAAS,WACT,cAAc,OACd,iBACA,SAAS,OACV,EACD,SAAS,aACTC,iCAAiB,SAAS,EAC1B,wBACA,+BACD;EACD,MAAM,eAAeC,8CAAyB,QAAQ;AAUtD,MAAI,CATc,MAAM,WACtB,IACA,QACA,SACA,cAAc,QACd,cAAc,MACd,QAAQ,iBACR,QAAQ,eAAe,SAAS,YACjC,EACe;AACd,MAAG,SAAS;AACZ,gBAAa,SAAS,cAAc;AACpC,gBAAa,SAAS,kBAAkB,cAAc,QAAQ;;AAEhE,gBAAc,SAAS;AACvB;;AAIF,KAAIG,mCAAmB,SAAS,EAAE;EAChC,MAAM,eAAe,QAAQ,IAAI;GAC/B,QAAQ;GACR,MAAM;GACN,SAASX,+BAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE,QAAQ;IAAK;IAAS;GACnC,CAAC;EACF,MAAM,SAASY,4CACb,SAAS,WACT,cAAc,OACd,WAIAN,yCACE,SAAS,WACT,cAAc,OACd,iBACA,SAAS,OACV,EACD,SAAS,aACTC,iCAAiB,SAAS,EAC1B,wBACA,+BACD;EACD,MAAM,eAAeC,8CAAyB,QAAQ;AAUtD,MAAI,CATc,MAAM,WACtB,IACA,QACA,SACA,cAAc,QACd,cAAc,MACd,QAAQ,iBACR,QAAQ,eAAe,SAAS,YACjC,EACe;AACd,MAAG,SAAS;AACZ,gBAAa,SAAS,cAAc;AACpC,gBAAa,SAAS,kBAAkB,cAAc,QAAQ;;AAEhE,gBAAc,SAAS;AACvB;;AAIF,SAAQ,IAAI;EACV,QAAQ;EACR,MAAM;EACN,SAASR,+BAAe,SAAS,kBAAkB,EAAE,CAAC;EACtD,MAAM;EACN,UAAU;GAAE,QAAQ;GAAK;GAAS;EACnC,CAAC;AACF,IAAG,KACD,KAAK,UACH,gBAAgB,iDAAiD,eAAe,CACjF,CACF;;AAGH,eAAe,WACb,IACA,QACA,SACA,QACA,aACA,iBACA,aACkB;CAClB,IAAI,aAAa;AACjB,MAAK,MAAM,SAAS,QAAQ;AAC1B,MAAI,GAAG,SAAU,QAAO;EACxB,MAAM,aAAaa,kCAAe,YAAY,QAAW,SAAS,iBAAiB,YAAY;AAC/F,MAAI,aAAa,EAAG,OAAMC,yBAAM,YAAY,OAAO;AACnD,MAAI,QAAQ,QAAS,QAAO;AAC5B,MAAI,GAAG,SAAU,QAAO;AACxB,KAAG,KAAK,KAAK,UAAU,MAAM,CAAC;AAC9B;AACA,iBAAe;AACf,MAAI,QAAQ,QAAS,QAAO;;AAE9B,QAAO"}

package/dist/ws-responses.d.cts.map

@@ -1 +1 @@
-{"version":3,"file":"ws-responses.d.cts","names":[],"sources":["../src/ws-responses.ts"],"sourcesContent":[],"mappings":";;;;;;;;AAwEW,iBAHK,wBAAA,CAGL,EAAA,EAFL,mBAEK,EAAA,QAAA,EADC,OACD,EAAA,EAAA,OAAA,EAAA,OAAA,EAAA,QAAA,EAAA;SAMC,EAAA,MAAA;WAEiB,EAAA,MAAA;aAA0B,CAAA,EAAA,MAAA;OAAqB,EAAA,MAAA;EAEhB,MAAA,EAJhD,MAIgD;;2BAF/B,0BAA0B;;mBAAqB,UAAA,CAEnC"}
+{"version":3,"file":"ws-responses.d.cts","names":[],"sources":["../src/ws-responses.ts"],"sourcesContent":[],"mappings":";;;;;;;;AA0EW,iBAHK,wBAAA,CAGL,EAAA,EAFL,mBAEK,EAAA,QAAA,EADC,OACD,EAAA,EAAA,OAAA,EAAA,OAAA,EAAA,QAAA,EAAA;SAMC,EAAA,MAAA;WAEiB,EAAA,MAAA;aAA0B,CAAA,EAAA,MAAA;OAAqB,EAAA,MAAA;EAEhB,MAAA,EAJhD,MAIgD;;2BAF/B,0BAA0B;;mBAAqB,UAAA,CAEnC"}

package/dist/ws-responses.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"ws-responses.d.ts","names":[],"sources":["../src/ws-responses.ts"],"sourcesContent":[],"mappings":";;;;;;;;AAwEW,iBAHK,wBAAA,CAGL,EAAA,EAFL,mBAEK,EAAA,QAAA,EADC,OACD,EAAA,EAAA,OAAA,EAAA,OAAA,EAAA,QAAA,EAAA;SAMC,EAAA,MAAA;WAEiB,EAAA,MAAA;aAA0B,CAAA,EAAA,MAAA;OAAqB,EAAA,MAAA;EAEhB,MAAA,EAJhD,MAIgD;;2BAF/B,0BAA0B;;mBAAqB,UAAA,CAEnC"}
+{"version":3,"file":"ws-responses.d.ts","names":[],"sources":["../src/ws-responses.ts"],"sourcesContent":[],"mappings":";;;;;;;;AA0EW,iBAHK,wBAAA,CAGL,EAAA,EAFL,mBAEK,EAAA,QAAA,EADC,OACD,EAAA,EAAA,OAAA,EAAA,OAAA,EAAA,QAAA,EAAA;SAMC,EAAA,MAAA;WAEiB,EAAA,MAAA;aAA0B,CAAA,EAAA,MAAA;OAAqB,EAAA,MAAA;EAEhB,MAAA,EAJhD,MAIgD;;2BAF/B,0BAA0B;;mBAAqB,UAAA,CAEnC"}

package/dist/ws-responses.js

@@ -4,7 +4,7 @@ import "./journal.js";
 import { matchFixtureDiagnostic } from "./router.js";
 import { calculateDelay, delay } from "./sse-writer.js";
 import { createInterruptionSignal } from "./interruption.js";
-import { buildContentWithToolCallsStreamEvents, buildTextStreamEvents, buildToolCallStreamEvents, responsesToCompletionRequest } from "./responses.js";
+import { buildContentWithToolCallsStreamEvents, buildTextStreamEvents, buildToolCallStreamEvents, requestIncludesEncryptedReasoning, requestWantsEncryptedReasoning, responsesToCompletionRequest } from "./responses.js";
 
 //#region src/ws-responses.ts
 function isResponseCreateMessage(msg) {
@@ -48,7 +48,7 @@ async function processMessage(raw, ws, fixtures, journal, defaults) {
 		ws.send(JSON.stringify(buildErrorEvent("Expected message type \"response.create\"", "invalid_request_error", "invalid_message_type")));
 		return;
 	}
-	const completionReq = responsesToCompletionRequest({
+	const responsesReq = {
 		model: parsed.model ?? defaults.model,
 		input: parsed.input ?? [],
 		instructions: parsed.instructions,
@@ -56,8 +56,13 @@ async function processMessage(raw, ws, fixtures, journal, defaults) {
 		tool_choice: parsed.tool_choice,
 		stream: parsed.stream,
 		temperature: parsed.temperature,
-		max_output_tokens: parsed.max_output_tokens
-	});
+		max_output_tokens: parsed.max_output_tokens,
+		include: parsed.include,
+		store: parsed.store
+	};
+	const emitEncryptedReasoning = requestWantsEncryptedReasoning(responsesReq);
+	const synthesizeSummarylessReasoning = requestIncludesEncryptedReasoning(responsesReq);
+	const completionReq = responsesToCompletionRequest(responsesReq);
 	completionReq._endpointType = "chat";
 	const contextHeader = defaults.upgradeHeaders?.["x-aimock-context"];
 	completionReq._context = typeof contextHeader === "string" ? contextHeader : Array.isArray(contextHeader) && contextHeader.length > 0 ? contextHeader[0] : void 0;
@@ -126,7 +131,7 @@ async function processMessage(raw, ws, fixtures, journal, defaults) {
 				fixture
 			}
 		});
-		const events = buildContentWithToolCallsStreamEvents(response.content ?? "", response.toolCalls ?? [], completionReq.model, chunkSize, resolveReasoningForModel(response.reasoning, completionReq.model, effectiveStrict, defaults.logger), response.webSearches, extractOverrides(response), response.blocks);
+		const events = buildContentWithToolCallsStreamEvents(response.content ?? "", response.toolCalls ?? [], completionReq.model, chunkSize, resolveReasoningForModel(response.reasoning, completionReq.model, effectiveStrict, defaults.logger), response.webSearches, extractOverrides(response), response.blocks, emitEncryptedReasoning, synthesizeSummarylessReasoning);
 		const interruption = createInterruptionSignal(fixture);
 		if (!await sendEvents(ws, events, latency, interruption?.signal, interruption?.tick, fixture.recordedTimings, fixture.replaySpeed ?? defaults.replaySpeed)) {
 			ws.destroy();
@@ -147,7 +152,7 @@ async function processMessage(raw, ws, fixtures, journal, defaults) {
 				fixture
 			}
 		});
-		const events = buildTextStreamEvents(response.content, completionReq.model, chunkSize, resolveReasoningForModel(response.reasoning, completionReq.model, effectiveStrict, defaults.logger), response.webSearches, extractOverrides(response));
+		const events = buildTextStreamEvents(response.content, completionReq.model, chunkSize, resolveReasoningForModel(response.reasoning, completionReq.model, effectiveStrict, defaults.logger), response.webSearches, extractOverrides(response), emitEncryptedReasoning, synthesizeSummarylessReasoning);
 		const interruption = createInterruptionSignal(fixture);
 		if (!await sendEvents(ws, events, latency, interruption?.signal, interruption?.tick, fixture.recordedTimings, fixture.replaySpeed ?? defaults.replaySpeed)) {
 			ws.destroy();
@@ -168,7 +173,7 @@ async function processMessage(raw, ws, fixtures, journal, defaults) {
 				fixture
 			}
 		});
-		const events = buildToolCallStreamEvents(response.toolCalls, completionReq.model, chunkSize, resolveReasoningForModel(response.reasoning, completionReq.model, effectiveStrict, defaults.logger), response.webSearches, extractOverrides(response));
+		const events = buildToolCallStreamEvents(response.toolCalls, completionReq.model, chunkSize, resolveReasoningForModel(response.reasoning, completionReq.model, effectiveStrict, defaults.logger), response.webSearches, extractOverrides(response), emitEncryptedReasoning, synthesizeSummarylessReasoning);
 		const interruption = createInterruptionSignal(fixture);
 		if (!await sendEvents(ws, events, latency, interruption?.signal, interruption?.tick, fixture.recordedTimings, fixture.replaySpeed ?? defaults.replaySpeed)) {
 			ws.destroy();

package/dist/ws-responses.js.map

@@ -1 +1 @@
-{"version":3,"file":"ws-responses.js","names":[],"sources":["../src/ws-responses.ts"],"sourcesContent":["/**\n * WebSocket handler for OpenAI Responses API.\n *\n * Accepts `{ type: \"response.create\", model: \"...\", input: [...] }` messages over\n * WebSocket and sends back the same Responses API SSE events as the HTTP\n * handler, but as individual WebSocket text frames.\n */\n\nimport type { ChatCompletionRequest, Fixture } from \"./types.js\";\nimport { matchFixtureDiagnostic } from \"./router.js\";\nimport {\n  responsesToCompletionRequest,\n  buildTextStreamEvents,\n  buildToolCallStreamEvents,\n  buildContentWithToolCallsStreamEvents,\n  type ResponsesSSEEvent,\n} from \"./responses.js\";\nimport {\n  isTextResponse,\n  isToolCallResponse,\n  isContentWithToolCallsResponse,\n  isErrorResponse,\n  extractOverrides,\n  resolveResponse,\n  resolveStrictMode,\n  resolveReasoningForModel,\n  strictOverrideField,\n  flattenHeaders,\n  strictNoMatchMessage,\n  strictNoMatchLogLine,\n} from \"./helpers.js\";\nimport { createInterruptionSignal } from \"./interruption.js\";\nimport { delay, calculateDelay } from \"./sse-writer.js\";\nimport { DEFAULT_TEST_ID, type Journal } from \"./journal.js\";\nimport type { Logger } from \"./logger.js\";\nimport type { WebSocketConnection } from \"./ws-framing.js\";\n\ninterface ResponseCreateMessage {\n  type: \"response.create\";\n  model?: string;\n  input?: unknown[];\n  instructions?: string;\n  tools?: unknown[];\n  tool_choice?: string | object;\n  stream?: boolean;\n  temperature?: number;\n  max_output_tokens?: number;\n  [key: string]: unknown;\n}\n\nfunction isResponseCreateMessage(msg: unknown): msg is ResponseCreateMessage {\n  return (\n    typeof msg === \"object\" &&\n    msg !== null &&\n    (msg as ResponseCreateMessage).type === \"response.create\"\n  );\n}\n\nfunction buildErrorEvent(\n  message: string,\n  type = \"invalid_request_error\",\n  code?: string,\n): ResponsesSSEEvent {\n  return {\n    type: \"error\",\n    error: { message, type, code },\n  };\n}\n\nexport function handleWebSocketResponses(\n  ws: WebSocketConnection,\n  fixtures: Fixture[],\n  journal: Journal,\n  defaults: {\n    latency: number;\n    chunkSize: number;\n    replaySpeed?: number;\n    model: string;\n    logger: Logger;\n    strict?: boolean;\n    requestTransform?: (req: ChatCompletionRequest) => ChatCompletionRequest;\n    testId?: string;\n    upgradeHeaders?: import(\"node:http\").IncomingHttpHeaders;\n  },\n): void {\n  const { logger } = defaults;\n  // Serialize message processing to prevent event interleaving\n  let pending = Promise.resolve();\n  ws.on(\"message\", (raw: string) => {\n    pending = pending.then(() =>\n      processMessage(raw, ws, fixtures, journal, defaults).catch((err: unknown) => {\n        const msg = err instanceof Error ? err.message : \"Internal error\";\n        logger.error(`WebSocket responses error: ${msg}`);\n        try {\n          ws.send(JSON.stringify(buildErrorEvent(msg, \"server_error\")));\n        } catch (sendErr) {\n          defaults.logger.debug(\n            `Failed to send error to client: ${sendErr instanceof Error ? sendErr.message : \"unknown\"}`,\n          );\n        }\n      }),\n    );\n  });\n}\n\nasync function processMessage(\n  raw: string,\n  ws: WebSocketConnection,\n  fixtures: Fixture[],\n  journal: Journal,\n  defaults: {\n    latency: number;\n    chunkSize: number;\n    replaySpeed?: number;\n    model: string;\n    logger: Logger;\n    strict?: boolean;\n    requestTransform?: (req: ChatCompletionRequest) => ChatCompletionRequest;\n    testId?: string;\n    upgradeHeaders?: import(\"node:http\").IncomingHttpHeaders;\n  },\n): Promise<void> {\n  let parsed: unknown;\n  try {\n    parsed = JSON.parse(raw);\n  } catch (parseErr) {\n    const detail = parseErr instanceof Error ? parseErr.message : \"unknown\";\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(`Malformed JSON: ${detail}`, \"invalid_request_error\", \"invalid_json\"),\n      ),\n    );\n    return;\n  }\n\n  if (!isResponseCreateMessage(parsed)) {\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(\n          'Expected message type \"response.create\"',\n          \"invalid_request_error\",\n          \"invalid_message_type\",\n        ),\n      ),\n    );\n    return;\n  }\n\n  const responsesReq = {\n    model: parsed.model ?? defaults.model,\n    input: (parsed.input ?? []) as {\n      role?: string;\n      type?: string;\n      content?: string | { type: string; text?: string }[];\n      call_id?: string;\n      name?: string;\n      arguments?: string;\n      output?: string;\n      id?: string;\n    }[],\n    instructions: parsed.instructions,\n    tools: parsed.tools as\n      | {\n          type: \"function\";\n          name: string;\n          description?: string;\n          parameters?: object;\n          strict?: boolean;\n        }[]\n      | undefined,\n    tool_choice: parsed.tool_choice,\n    stream: parsed.stream,\n    temperature: parsed.temperature,\n    max_output_tokens: parsed.max_output_tokens,\n  };\n\n  const completionReq = responsesToCompletionRequest(responsesReq);\n  completionReq._endpointType = \"chat\";\n  const contextHeader = defaults.upgradeHeaders?.[\"x-aimock-context\"];\n  completionReq._context =\n    typeof contextHeader === \"string\"\n      ? contextHeader\n      : Array.isArray(contextHeader) && contextHeader.length > 0\n        ? contextHeader[0]\n        : undefined;\n  const testId = defaults.testId ?? DEFAULT_TEST_ID;\n  const { fixture, skippedBySequenceOrTurn } = matchFixtureDiagnostic(\n    fixtures,\n    completionReq,\n    journal.getFixtureMatchCountsForTest(testId),\n    defaults.requestTransform,\n  );\n\n  if (fixture) {\n    journal.incrementFixtureMatchCount(fixture, fixtures, testId);\n  }\n\n  if (!fixture) {\n    if (resolveStrictMode(defaults.strict, defaults.upgradeHeaders)) {\n      const strictMessage = strictNoMatchMessage(skippedBySequenceOrTurn);\n      defaults.logger.error(strictNoMatchLogLine(\"WS\", \"/v1/responses\", skippedBySequenceOrTurn));\n      journal.add({\n        method: \"WS\",\n        path: \"/v1/responses\",\n        headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n        body: completionReq,\n        response: {\n          status: 503,\n          fixture: null,\n          ...strictOverrideField(defaults.strict, defaults.upgradeHeaders),\n        },\n      });\n      ws.close(1008, strictMessage);\n      return;\n    }\n    journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: {\n        status: 404,\n        fixture: null,\n        ...strictOverrideField(defaults.strict, defaults.upgradeHeaders),\n      },\n    });\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(\"No fixture matched\", \"invalid_request_error\", \"no_fixture_match\"),\n      ),\n    );\n    return;\n  }\n\n  const response = await resolveResponse(fixture, completionReq);\n  const latency = fixture.latency ?? defaults.latency;\n  const chunkSize = Math.max(1, fixture.chunkSize ?? defaults.chunkSize);\n\n  // The WS path has no per-request `req.headers`; strict is resolved from the\n  // connection's upgrade headers (see the `!fixture` branch above). Used below\n  // to gate the synthesized reasoning channel on the requested model's capability.\n  const effectiveStrict = resolveStrictMode(defaults.strict, defaults.upgradeHeaders);\n\n  // Error response\n  if (isErrorResponse(response)) {\n    const status = response.status ?? 500;\n    journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status, fixture },\n    });\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(response.error.message, response.error.type, response.error.code),\n      ),\n    );\n    return;\n  }\n\n  // Content + tool calls response (must be checked before isTextResponse / isToolCallResponse)\n  if (isContentWithToolCallsResponse(response)) {\n    const journalEntry = journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status: 200, fixture },\n    });\n\n    const events = buildContentWithToolCallsStreamEvents(\n      response.content ?? \"\",\n      response.toolCalls ?? [],\n      completionReq.model,\n      chunkSize,\n      resolveReasoningForModel(\n        response.reasoning,\n        completionReq.model,\n        effectiveStrict,\n        defaults.logger,\n      ),\n      response.webSearches,\n      extractOverrides(response),\n      response.blocks,\n    );\n\n    const interruption = createInterruptionSignal(fixture);\n    const completed = await sendEvents(\n      ws,\n      events,\n      latency,\n      interruption?.signal,\n      interruption?.tick,\n      fixture.recordedTimings,\n      fixture.replaySpeed ?? defaults.replaySpeed,\n    );\n    if (!completed) {\n      ws.destroy();\n      journalEntry.response.interrupted = true;\n      journalEntry.response.interruptReason = interruption?.reason();\n    }\n    interruption?.cleanup();\n    return;\n  }\n\n  // Text response\n  if (isTextResponse(response)) {\n    const journalEntry = journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status: 200, fixture },\n    });\n\n    const events = buildTextStreamEvents(\n      response.content,\n      completionReq.model,\n      chunkSize,\n      resolveReasoningForModel(\n        response.reasoning,\n        completionReq.model,\n        effectiveStrict,\n        defaults.logger,\n      ),\n      response.webSearches,\n      extractOverrides(response),\n    );\n    const interruption = createInterruptionSignal(fixture);\n    const completed = await sendEvents(\n      ws,\n      events,\n      latency,\n      interruption?.signal,\n      interruption?.tick,\n      fixture.recordedTimings,\n      fixture.replaySpeed ?? defaults.replaySpeed,\n    );\n    if (!completed) {\n      ws.destroy();\n      journalEntry.response.interrupted = true;\n      journalEntry.response.interruptReason = interruption?.reason();\n    }\n    interruption?.cleanup();\n    return;\n  }\n\n  // Tool call response\n  if (isToolCallResponse(response)) {\n    const journalEntry = journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status: 200, fixture },\n    });\n    const events = buildToolCallStreamEvents(\n      response.toolCalls,\n      completionReq.model,\n      chunkSize,\n      // Gate the synthesized reasoning channel on the requested model's\n      // capability, matching the WS text / content+tool branches and the HTTP\n      // tool-only path so reasoning emission is transport-independent.\n      resolveReasoningForModel(\n        response.reasoning,\n        completionReq.model,\n        effectiveStrict,\n        defaults.logger,\n      ),\n      response.webSearches,\n      extractOverrides(response),\n    );\n    const interruption = createInterruptionSignal(fixture);\n    const completed = await sendEvents(\n      ws,\n      events,\n      latency,\n      interruption?.signal,\n      interruption?.tick,\n      fixture.recordedTimings,\n      fixture.replaySpeed ?? defaults.replaySpeed,\n    );\n    if (!completed) {\n      ws.destroy();\n      journalEntry.response.interrupted = true;\n      journalEntry.response.interruptReason = interruption?.reason();\n    }\n    interruption?.cleanup();\n    return;\n  }\n\n  // Unknown response type\n  journal.add({\n    method: \"WS\",\n    path: \"/v1/responses\",\n    headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n    body: completionReq,\n    response: { status: 500, fixture },\n  });\n  ws.send(\n    JSON.stringify(\n      buildErrorEvent(\"Fixture response did not match any known type\", \"server_error\"),\n    ),\n  );\n}\n\nasync function sendEvents(\n  ws: WebSocketConnection,\n  events: ResponsesSSEEvent[],\n  latency: number,\n  signal?: AbortSignal,\n  onChunkSent?: () => void,\n  recordedTimings?: import(\"./types.js\").RecordedTimings,\n  replaySpeed?: number,\n): Promise<boolean> {\n  let eventIndex = 0;\n  for (const event of events) {\n    if (ws.isClosed) return false;\n    const chunkDelay = calculateDelay(eventIndex, undefined, latency, recordedTimings, replaySpeed);\n    if (chunkDelay > 0) await delay(chunkDelay, signal);\n    if (signal?.aborted) return false;\n    if (ws.isClosed) return false;\n    ws.send(JSON.stringify(event));\n    eventIndex++;\n    onChunkSent?.();\n    if (signal?.aborted) return false;\n  }\n  return true;\n}\n"],"mappings":";;;;;;;;;AAkDA,SAAS,wBAAwB,KAA4C;AAC3E,QACE,OAAO,QAAQ,YACf,QAAQ,QACP,IAA8B,SAAS;;AAI5C,SAAS,gBACP,SACA,OAAO,yBACP,MACmB;AACnB,QAAO;EACL,MAAM;EACN,OAAO;GAAE;GAAS;GAAM;GAAM;EAC/B;;AAGH,SAAgB,yBACd,IACA,UACA,SACA,UAWM;CACN,MAAM,EAAE,WAAW;CAEnB,IAAI,UAAU,QAAQ,SAAS;AAC/B,IAAG,GAAG,YAAY,QAAgB;AAChC,YAAU,QAAQ,WAChB,eAAe,KAAK,IAAI,UAAU,SAAS,SAAS,CAAC,OAAO,QAAiB;GAC3E,MAAM,MAAM,eAAe,QAAQ,IAAI,UAAU;AACjD,UAAO,MAAM,8BAA8B,MAAM;AACjD,OAAI;AACF,OAAG,KAAK,KAAK,UAAU,gBAAgB,KAAK,eAAe,CAAC,CAAC;YACtD,SAAS;AAChB,aAAS,OAAO,MACd,mCAAmC,mBAAmB,QAAQ,QAAQ,UAAU,YACjF;;IAEH,CACH;GACD;;AAGJ,eAAe,eACb,KACA,IACA,UACA,SACA,UAWe;CACf,IAAI;AACJ,KAAI;AACF,WAAS,KAAK,MAAM,IAAI;UACjB,UAAU;EACjB,MAAM,SAAS,oBAAoB,QAAQ,SAAS,UAAU;AAC9D,KAAG,KACD,KAAK,UACH,gBAAgB,mBAAmB,UAAU,yBAAyB,eAAe,CACtF,CACF;AACD;;AAGF,KAAI,CAAC,wBAAwB,OAAO,EAAE;AACpC,KAAG,KACD,KAAK,UACH,gBACE,6CACA,yBACA,uBACD,CACF,CACF;AACD;;CA+BF,MAAM,gBAAgB,6BA5BD;EACnB,OAAO,OAAO,SAAS,SAAS;EAChC,OAAQ,OAAO,SAAS,EAAE;EAU1B,cAAc,OAAO;EACrB,OAAO,OAAO;EASd,aAAa,OAAO;EACpB,QAAQ,OAAO;EACf,aAAa,OAAO;EACpB,mBAAmB,OAAO;EAC3B,CAE+D;AAChE,eAAc,gBAAgB;CAC9B,MAAM,gBAAgB,SAAS,iBAAiB;AAChD,eAAc,WACZ,OAAO,kBAAkB,WACrB,gBACA,MAAM,QAAQ,cAAc,IAAI,cAAc,SAAS,IACrD,cAAc,KACd;CACR,MAAM,SAAS,SAAS,UAAU;CAClC,MAAM,EAAE,SAAS,4BAA4B,uBAC3C,UACA,eACA,QAAQ,6BAA6B,OAAO,EAC5C,SAAS,iBACV;AAED,KAAI,QACF,SAAQ,2BAA2B,SAAS,UAAU,OAAO;AAG/D,KAAI,CAAC,SAAS;AACZ,MAAI,kBAAkB,SAAS,QAAQ,SAAS,eAAe,EAAE;GAC/D,MAAM,gBAAgB,qBAAqB,wBAAwB;AACnE,YAAS,OAAO,MAAM,qBAAqB,MAAM,iBAAiB,wBAAwB,CAAC;AAC3F,WAAQ,IAAI;IACV,QAAQ;IACR,MAAM;IACN,SAAS,eAAe,SAAS,kBAAkB,EAAE,CAAC;IACtD,MAAM;IACN,UAAU;KACR,QAAQ;KACR,SAAS;KACT,GAAG,oBAAoB,SAAS,QAAQ,SAAS,eAAe;KACjE;IACF,CAAC;AACF,MAAG,MAAM,MAAM,cAAc;AAC7B;;AAEF,UAAQ,IAAI;GACV,QAAQ;GACR,MAAM;GACN,SAAS,eAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IACR,QAAQ;IACR,SAAS;IACT,GAAG,oBAAoB,SAAS,QAAQ,SAAS,eAAe;IACjE;GACF,CAAC;AACF,KAAG,KACD,KAAK,UACH,gBAAgB,sBAAsB,yBAAyB,mBAAmB,CACnF,CACF;AACD;;CAGF,MAAM,WAAW,MAAM,gBAAgB,SAAS,cAAc;CAC9D,MAAM,UAAU,QAAQ,WAAW,SAAS;CAC5C,MAAM,YAAY,KAAK,IAAI,GAAG,QAAQ,aAAa,SAAS,UAAU;CAKtE,MAAM,kBAAkB,kBAAkB,SAAS,QAAQ,SAAS,eAAe;AAGnF,KAAI,gBAAgB,SAAS,EAAE;EAC7B,MAAM,SAAS,SAAS,UAAU;AAClC,UAAQ,IAAI;GACV,QAAQ;GACR,MAAM;GACN,SAAS,eAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE;IAAQ;IAAS;GAC9B,CAAC;AACF,KAAG,KACD,KAAK,UACH,gBAAgB,SAAS,MAAM,SAAS,SAAS,MAAM,MAAM,SAAS,MAAM,KAAK,CAClF,CACF;AACD;;AAIF,KAAI,+BAA+B,SAAS,EAAE;EAC5C,MAAM,eAAe,QAAQ,IAAI;GAC/B,QAAQ;GACR,MAAM;GACN,SAAS,eAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE,QAAQ;IAAK;IAAS;GACnC,CAAC;EAEF,MAAM,SAAS,sCACb,SAAS,WAAW,IACpB,SAAS,aAAa,EAAE,EACxB,cAAc,OACd,WACA,yBACE,SAAS,WACT,cAAc,OACd,iBACA,SAAS,OACV,EACD,SAAS,aACT,iBAAiB,SAAS,EAC1B,SAAS,OACV;EAED,MAAM,eAAe,yBAAyB,QAAQ;AAUtD,MAAI,CATc,MAAM,WACtB,IACA,QACA,SACA,cAAc,QACd,cAAc,MACd,QAAQ,iBACR,QAAQ,eAAe,SAAS,YACjC,EACe;AACd,MAAG,SAAS;AACZ,gBAAa,SAAS,cAAc;AACpC,gBAAa,SAAS,kBAAkB,cAAc,QAAQ;;AAEhE,gBAAc,SAAS;AACvB;;AAIF,KAAI,eAAe,SAAS,EAAE;EAC5B,MAAM,eAAe,QAAQ,IAAI;GAC/B,QAAQ;GACR,MAAM;GACN,SAAS,eAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE,QAAQ;IAAK;IAAS;GACnC,CAAC;EAEF,MAAM,SAAS,sBACb,SAAS,SACT,cAAc,OACd,WACA,yBACE,SAAS,WACT,cAAc,OACd,iBACA,SAAS,OACV,EACD,SAAS,aACT,iBAAiB,SAAS,CAC3B;EACD,MAAM,eAAe,yBAAyB,QAAQ;AAUtD,MAAI,CATc,MAAM,WACtB,IACA,QACA,SACA,cAAc,QACd,cAAc,MACd,QAAQ,iBACR,QAAQ,eAAe,SAAS,YACjC,EACe;AACd,MAAG,SAAS;AACZ,gBAAa,SAAS,cAAc;AACpC,gBAAa,SAAS,kBAAkB,cAAc,QAAQ;;AAEhE,gBAAc,SAAS;AACvB;;AAIF,KAAI,mBAAmB,SAAS,EAAE;EAChC,MAAM,eAAe,QAAQ,IAAI;GAC/B,QAAQ;GACR,MAAM;GACN,SAAS,eAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE,QAAQ;IAAK;IAAS;GACnC,CAAC;EACF,MAAM,SAAS,0BACb,SAAS,WACT,cAAc,OACd,WAIA,yBACE,SAAS,WACT,cAAc,OACd,iBACA,SAAS,OACV,EACD,SAAS,aACT,iBAAiB,SAAS,CAC3B;EACD,MAAM,eAAe,yBAAyB,QAAQ;AAUtD,MAAI,CATc,MAAM,WACtB,IACA,QACA,SACA,cAAc,QACd,cAAc,MACd,QAAQ,iBACR,QAAQ,eAAe,SAAS,YACjC,EACe;AACd,MAAG,SAAS;AACZ,gBAAa,SAAS,cAAc;AACpC,gBAAa,SAAS,kBAAkB,cAAc,QAAQ;;AAEhE,gBAAc,SAAS;AACvB;;AAIF,SAAQ,IAAI;EACV,QAAQ;EACR,MAAM;EACN,SAAS,eAAe,SAAS,kBAAkB,EAAE,CAAC;EACtD,MAAM;EACN,UAAU;GAAE,QAAQ;GAAK;GAAS;EACnC,CAAC;AACF,IAAG,KACD,KAAK,UACH,gBAAgB,iDAAiD,eAAe,CACjF,CACF;;AAGH,eAAe,WACb,IACA,QACA,SACA,QACA,aACA,iBACA,aACkB;CAClB,IAAI,aAAa;AACjB,MAAK,MAAM,SAAS,QAAQ;AAC1B,MAAI,GAAG,SAAU,QAAO;EACxB,MAAM,aAAa,eAAe,YAAY,QAAW,SAAS,iBAAiB,YAAY;AAC/F,MAAI,aAAa,EAAG,OAAM,MAAM,YAAY,OAAO;AACnD,MAAI,QAAQ,QAAS,QAAO;AAC5B,MAAI,GAAG,SAAU,QAAO;AACxB,KAAG,KAAK,KAAK,UAAU,MAAM,CAAC;AAC9B;AACA,iBAAe;AACf,MAAI,QAAQ,QAAS,QAAO;;AAE9B,QAAO"}
+{"version":3,"file":"ws-responses.js","names":[],"sources":["../src/ws-responses.ts"],"sourcesContent":["/**\n * WebSocket handler for OpenAI Responses API.\n *\n * Accepts `{ type: \"response.create\", model: \"...\", input: [...] }` messages over\n * WebSocket and sends back the same Responses API SSE events as the HTTP\n * handler, but as individual WebSocket text frames.\n */\n\nimport type { ChatCompletionRequest, Fixture } from \"./types.js\";\nimport { matchFixtureDiagnostic } from \"./router.js\";\nimport {\n  responsesToCompletionRequest,\n  requestIncludesEncryptedReasoning,\n  requestWantsEncryptedReasoning,\n  buildTextStreamEvents,\n  buildToolCallStreamEvents,\n  buildContentWithToolCallsStreamEvents,\n  type ResponsesSSEEvent,\n} from \"./responses.js\";\nimport {\n  isTextResponse,\n  isToolCallResponse,\n  isContentWithToolCallsResponse,\n  isErrorResponse,\n  extractOverrides,\n  resolveResponse,\n  resolveStrictMode,\n  resolveReasoningForModel,\n  strictOverrideField,\n  flattenHeaders,\n  strictNoMatchMessage,\n  strictNoMatchLogLine,\n} from \"./helpers.js\";\nimport { createInterruptionSignal } from \"./interruption.js\";\nimport { delay, calculateDelay } from \"./sse-writer.js\";\nimport { DEFAULT_TEST_ID, type Journal } from \"./journal.js\";\nimport type { Logger } from \"./logger.js\";\nimport type { WebSocketConnection } from \"./ws-framing.js\";\n\ninterface ResponseCreateMessage {\n  type: \"response.create\";\n  model?: string;\n  input?: unknown[];\n  instructions?: string;\n  tools?: unknown[];\n  tool_choice?: string | object;\n  stream?: boolean;\n  temperature?: number;\n  max_output_tokens?: number;\n  [key: string]: unknown;\n}\n\nfunction isResponseCreateMessage(msg: unknown): msg is ResponseCreateMessage {\n  return (\n    typeof msg === \"object\" &&\n    msg !== null &&\n    (msg as ResponseCreateMessage).type === \"response.create\"\n  );\n}\n\nfunction buildErrorEvent(\n  message: string,\n  type = \"invalid_request_error\",\n  code?: string,\n): ResponsesSSEEvent {\n  return {\n    type: \"error\",\n    error: { message, type, code },\n  };\n}\n\nexport function handleWebSocketResponses(\n  ws: WebSocketConnection,\n  fixtures: Fixture[],\n  journal: Journal,\n  defaults: {\n    latency: number;\n    chunkSize: number;\n    replaySpeed?: number;\n    model: string;\n    logger: Logger;\n    strict?: boolean;\n    requestTransform?: (req: ChatCompletionRequest) => ChatCompletionRequest;\n    testId?: string;\n    upgradeHeaders?: import(\"node:http\").IncomingHttpHeaders;\n  },\n): void {\n  const { logger } = defaults;\n  // Serialize message processing to prevent event interleaving\n  let pending = Promise.resolve();\n  ws.on(\"message\", (raw: string) => {\n    pending = pending.then(() =>\n      processMessage(raw, ws, fixtures, journal, defaults).catch((err: unknown) => {\n        const msg = err instanceof Error ? err.message : \"Internal error\";\n        logger.error(`WebSocket responses error: ${msg}`);\n        try {\n          ws.send(JSON.stringify(buildErrorEvent(msg, \"server_error\")));\n        } catch (sendErr) {\n          defaults.logger.debug(\n            `Failed to send error to client: ${sendErr instanceof Error ? sendErr.message : \"unknown\"}`,\n          );\n        }\n      }),\n    );\n  });\n}\n\nasync function processMessage(\n  raw: string,\n  ws: WebSocketConnection,\n  fixtures: Fixture[],\n  journal: Journal,\n  defaults: {\n    latency: number;\n    chunkSize: number;\n    replaySpeed?: number;\n    model: string;\n    logger: Logger;\n    strict?: boolean;\n    requestTransform?: (req: ChatCompletionRequest) => ChatCompletionRequest;\n    testId?: string;\n    upgradeHeaders?: import(\"node:http\").IncomingHttpHeaders;\n  },\n): Promise<void> {\n  let parsed: unknown;\n  try {\n    parsed = JSON.parse(raw);\n  } catch (parseErr) {\n    const detail = parseErr instanceof Error ? parseErr.message : \"unknown\";\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(`Malformed JSON: ${detail}`, \"invalid_request_error\", \"invalid_json\"),\n      ),\n    );\n    return;\n  }\n\n  if (!isResponseCreateMessage(parsed)) {\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(\n          'Expected message type \"response.create\"',\n          \"invalid_request_error\",\n          \"invalid_message_type\",\n        ),\n      ),\n    );\n    return;\n  }\n\n  const responsesReq = {\n    model: parsed.model ?? defaults.model,\n    input: (parsed.input ?? []) as {\n      role?: string;\n      type?: string;\n      content?: string | { type: string; text?: string }[];\n      call_id?: string;\n      name?: string;\n      arguments?: string;\n      output?: string;\n      id?: string;\n    }[],\n    instructions: parsed.instructions,\n    tools: parsed.tools as\n      | {\n          type: \"function\";\n          name: string;\n          description?: string;\n          parameters?: object;\n          strict?: boolean;\n        }[]\n      | undefined,\n    tool_choice: parsed.tool_choice,\n    stream: parsed.stream,\n    temperature: parsed.temperature,\n    max_output_tokens: parsed.max_output_tokens,\n    include: (parsed as { include?: string[] }).include,\n    store: (parsed as { store?: boolean }).store,\n  };\n\n  // Gate encrypted-reasoning emission identically to the HTTP transport so the\n  // agent-framework#7233 stateless-replay path works over WebSocket too.\n  const emitEncryptedReasoning = requestWantsEncryptedReasoning(responsesReq);\n  // Synthesizing a reasoning item the fixture never declared takes the NARROWER\n  // explicit-`include` gate, transport-identically to HTTP — a `store: false`\n  // request with a summary-less fixture still gets no reasoning item.\n  const synthesizeSummarylessReasoning = requestIncludesEncryptedReasoning(responsesReq);\n\n  const completionReq = responsesToCompletionRequest(responsesReq);\n  completionReq._endpointType = \"chat\";\n  const contextHeader = defaults.upgradeHeaders?.[\"x-aimock-context\"];\n  completionReq._context =\n    typeof contextHeader === \"string\"\n      ? contextHeader\n      : Array.isArray(contextHeader) && contextHeader.length > 0\n        ? contextHeader[0]\n        : undefined;\n  const testId = defaults.testId ?? DEFAULT_TEST_ID;\n  const { fixture, skippedBySequenceOrTurn } = matchFixtureDiagnostic(\n    fixtures,\n    completionReq,\n    journal.getFixtureMatchCountsForTest(testId),\n    defaults.requestTransform,\n  );\n\n  if (fixture) {\n    journal.incrementFixtureMatchCount(fixture, fixtures, testId);\n  }\n\n  if (!fixture) {\n    if (resolveStrictMode(defaults.strict, defaults.upgradeHeaders)) {\n      const strictMessage = strictNoMatchMessage(skippedBySequenceOrTurn);\n      defaults.logger.error(strictNoMatchLogLine(\"WS\", \"/v1/responses\", skippedBySequenceOrTurn));\n      journal.add({\n        method: \"WS\",\n        path: \"/v1/responses\",\n        headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n        body: completionReq,\n        response: {\n          status: 503,\n          fixture: null,\n          ...strictOverrideField(defaults.strict, defaults.upgradeHeaders),\n        },\n      });\n      ws.close(1008, strictMessage);\n      return;\n    }\n    journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: {\n        status: 404,\n        fixture: null,\n        ...strictOverrideField(defaults.strict, defaults.upgradeHeaders),\n      },\n    });\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(\"No fixture matched\", \"invalid_request_error\", \"no_fixture_match\"),\n      ),\n    );\n    return;\n  }\n\n  const response = await resolveResponse(fixture, completionReq);\n  const latency = fixture.latency ?? defaults.latency;\n  const chunkSize = Math.max(1, fixture.chunkSize ?? defaults.chunkSize);\n\n  // The WS path has no per-request `req.headers`; strict is resolved from the\n  // connection's upgrade headers (see the `!fixture` branch above). Used below\n  // to gate the synthesized reasoning channel on the requested model's capability.\n  const effectiveStrict = resolveStrictMode(defaults.strict, defaults.upgradeHeaders);\n\n  // Error response\n  if (isErrorResponse(response)) {\n    const status = response.status ?? 500;\n    journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status, fixture },\n    });\n    ws.send(\n      JSON.stringify(\n        buildErrorEvent(response.error.message, response.error.type, response.error.code),\n      ),\n    );\n    return;\n  }\n\n  // Content + tool calls response (must be checked before isTextResponse / isToolCallResponse)\n  if (isContentWithToolCallsResponse(response)) {\n    const journalEntry = journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status: 200, fixture },\n    });\n\n    const events = buildContentWithToolCallsStreamEvents(\n      response.content ?? \"\",\n      response.toolCalls ?? [],\n      completionReq.model,\n      chunkSize,\n      resolveReasoningForModel(\n        response.reasoning,\n        completionReq.model,\n        effectiveStrict,\n        defaults.logger,\n      ),\n      response.webSearches,\n      extractOverrides(response),\n      response.blocks,\n      emitEncryptedReasoning,\n      synthesizeSummarylessReasoning,\n    );\n\n    const interruption = createInterruptionSignal(fixture);\n    const completed = await sendEvents(\n      ws,\n      events,\n      latency,\n      interruption?.signal,\n      interruption?.tick,\n      fixture.recordedTimings,\n      fixture.replaySpeed ?? defaults.replaySpeed,\n    );\n    if (!completed) {\n      ws.destroy();\n      journalEntry.response.interrupted = true;\n      journalEntry.response.interruptReason = interruption?.reason();\n    }\n    interruption?.cleanup();\n    return;\n  }\n\n  // Text response\n  if (isTextResponse(response)) {\n    const journalEntry = journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status: 200, fixture },\n    });\n\n    const events = buildTextStreamEvents(\n      response.content,\n      completionReq.model,\n      chunkSize,\n      resolveReasoningForModel(\n        response.reasoning,\n        completionReq.model,\n        effectiveStrict,\n        defaults.logger,\n      ),\n      response.webSearches,\n      extractOverrides(response),\n      emitEncryptedReasoning,\n      synthesizeSummarylessReasoning,\n    );\n    const interruption = createInterruptionSignal(fixture);\n    const completed = await sendEvents(\n      ws,\n      events,\n      latency,\n      interruption?.signal,\n      interruption?.tick,\n      fixture.recordedTimings,\n      fixture.replaySpeed ?? defaults.replaySpeed,\n    );\n    if (!completed) {\n      ws.destroy();\n      journalEntry.response.interrupted = true;\n      journalEntry.response.interruptReason = interruption?.reason();\n    }\n    interruption?.cleanup();\n    return;\n  }\n\n  // Tool call response\n  if (isToolCallResponse(response)) {\n    const journalEntry = journal.add({\n      method: \"WS\",\n      path: \"/v1/responses\",\n      headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n      body: completionReq,\n      response: { status: 200, fixture },\n    });\n    const events = buildToolCallStreamEvents(\n      response.toolCalls,\n      completionReq.model,\n      chunkSize,\n      // Gate the synthesized reasoning channel on the requested model's\n      // capability, matching the WS text / content+tool branches and the HTTP\n      // tool-only path so reasoning emission is transport-independent.\n      resolveReasoningForModel(\n        response.reasoning,\n        completionReq.model,\n        effectiveStrict,\n        defaults.logger,\n      ),\n      response.webSearches,\n      extractOverrides(response),\n      emitEncryptedReasoning,\n      synthesizeSummarylessReasoning,\n    );\n    const interruption = createInterruptionSignal(fixture);\n    const completed = await sendEvents(\n      ws,\n      events,\n      latency,\n      interruption?.signal,\n      interruption?.tick,\n      fixture.recordedTimings,\n      fixture.replaySpeed ?? defaults.replaySpeed,\n    );\n    if (!completed) {\n      ws.destroy();\n      journalEntry.response.interrupted = true;\n      journalEntry.response.interruptReason = interruption?.reason();\n    }\n    interruption?.cleanup();\n    return;\n  }\n\n  // Unknown response type\n  journal.add({\n    method: \"WS\",\n    path: \"/v1/responses\",\n    headers: flattenHeaders(defaults.upgradeHeaders ?? {}),\n    body: completionReq,\n    response: { status: 500, fixture },\n  });\n  ws.send(\n    JSON.stringify(\n      buildErrorEvent(\"Fixture response did not match any known type\", \"server_error\"),\n    ),\n  );\n}\n\nasync function sendEvents(\n  ws: WebSocketConnection,\n  events: ResponsesSSEEvent[],\n  latency: number,\n  signal?: AbortSignal,\n  onChunkSent?: () => void,\n  recordedTimings?: import(\"./types.js\").RecordedTimings,\n  replaySpeed?: number,\n): Promise<boolean> {\n  let eventIndex = 0;\n  for (const event of events) {\n    if (ws.isClosed) return false;\n    const chunkDelay = calculateDelay(eventIndex, undefined, latency, recordedTimings, replaySpeed);\n    if (chunkDelay > 0) await delay(chunkDelay, signal);\n    if (signal?.aborted) return false;\n    if (ws.isClosed) return false;\n    ws.send(JSON.stringify(event));\n    eventIndex++;\n    onChunkSent?.();\n    if (signal?.aborted) return false;\n  }\n  return true;\n}\n"],"mappings":";;;;;;;;;AAoDA,SAAS,wBAAwB,KAA4C;AAC3E,QACE,OAAO,QAAQ,YACf,QAAQ,QACP,IAA8B,SAAS;;AAI5C,SAAS,gBACP,SACA,OAAO,yBACP,MACmB;AACnB,QAAO;EACL,MAAM;EACN,OAAO;GAAE;GAAS;GAAM;GAAM;EAC/B;;AAGH,SAAgB,yBACd,IACA,UACA,SACA,UAWM;CACN,MAAM,EAAE,WAAW;CAEnB,IAAI,UAAU,QAAQ,SAAS;AAC/B,IAAG,GAAG,YAAY,QAAgB;AAChC,YAAU,QAAQ,WAChB,eAAe,KAAK,IAAI,UAAU,SAAS,SAAS,CAAC,OAAO,QAAiB;GAC3E,MAAM,MAAM,eAAe,QAAQ,IAAI,UAAU;AACjD,UAAO,MAAM,8BAA8B,MAAM;AACjD,OAAI;AACF,OAAG,KAAK,KAAK,UAAU,gBAAgB,KAAK,eAAe,CAAC,CAAC;YACtD,SAAS;AAChB,aAAS,OAAO,MACd,mCAAmC,mBAAmB,QAAQ,QAAQ,UAAU,YACjF;;IAEH,CACH;GACD;;AAGJ,eAAe,eACb,KACA,IACA,UACA,SACA,UAWe;CACf,IAAI;AACJ,KAAI;AACF,WAAS,KAAK,MAAM,IAAI;UACjB,UAAU;EACjB,MAAM,SAAS,oBAAoB,QAAQ,SAAS,UAAU;AAC9D,KAAG,KACD,KAAK,UACH,gBAAgB,mBAAmB,UAAU,yBAAyB,eAAe,CACtF,CACF;AACD;;AAGF,KAAI,CAAC,wBAAwB,OAAO,EAAE;AACpC,KAAG,KACD,KAAK,UACH,gBACE,6CACA,yBACA,uBACD,CACF,CACF;AACD;;CAGF,MAAM,eAAe;EACnB,OAAO,OAAO,SAAS,SAAS;EAChC,OAAQ,OAAO,SAAS,EAAE;EAU1B,cAAc,OAAO;EACrB,OAAO,OAAO;EASd,aAAa,OAAO;EACpB,QAAQ,OAAO;EACf,aAAa,OAAO;EACpB,mBAAmB,OAAO;EAC1B,SAAU,OAAkC;EAC5C,OAAQ,OAA+B;EACxC;CAID,MAAM,yBAAyB,+BAA+B,aAAa;CAI3E,MAAM,iCAAiC,kCAAkC,aAAa;CAEtF,MAAM,gBAAgB,6BAA6B,aAAa;AAChE,eAAc,gBAAgB;CAC9B,MAAM,gBAAgB,SAAS,iBAAiB;AAChD,eAAc,WACZ,OAAO,kBAAkB,WACrB,gBACA,MAAM,QAAQ,cAAc,IAAI,cAAc,SAAS,IACrD,cAAc,KACd;CACR,MAAM,SAAS,SAAS,UAAU;CAClC,MAAM,EAAE,SAAS,4BAA4B,uBAC3C,UACA,eACA,QAAQ,6BAA6B,OAAO,EAC5C,SAAS,iBACV;AAED,KAAI,QACF,SAAQ,2BAA2B,SAAS,UAAU,OAAO;AAG/D,KAAI,CAAC,SAAS;AACZ,MAAI,kBAAkB,SAAS,QAAQ,SAAS,eAAe,EAAE;GAC/D,MAAM,gBAAgB,qBAAqB,wBAAwB;AACnE,YAAS,OAAO,MAAM,qBAAqB,MAAM,iBAAiB,wBAAwB,CAAC;AAC3F,WAAQ,IAAI;IACV,QAAQ;IACR,MAAM;IACN,SAAS,eAAe,SAAS,kBAAkB,EAAE,CAAC;IACtD,MAAM;IACN,UAAU;KACR,QAAQ;KACR,SAAS;KACT,GAAG,oBAAoB,SAAS,QAAQ,SAAS,eAAe;KACjE;IACF,CAAC;AACF,MAAG,MAAM,MAAM,cAAc;AAC7B;;AAEF,UAAQ,IAAI;GACV,QAAQ;GACR,MAAM;GACN,SAAS,eAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IACR,QAAQ;IACR,SAAS;IACT,GAAG,oBAAoB,SAAS,QAAQ,SAAS,eAAe;IACjE;GACF,CAAC;AACF,KAAG,KACD,KAAK,UACH,gBAAgB,sBAAsB,yBAAyB,mBAAmB,CACnF,CACF;AACD;;CAGF,MAAM,WAAW,MAAM,gBAAgB,SAAS,cAAc;CAC9D,MAAM,UAAU,QAAQ,WAAW,SAAS;CAC5C,MAAM,YAAY,KAAK,IAAI,GAAG,QAAQ,aAAa,SAAS,UAAU;CAKtE,MAAM,kBAAkB,kBAAkB,SAAS,QAAQ,SAAS,eAAe;AAGnF,KAAI,gBAAgB,SAAS,EAAE;EAC7B,MAAM,SAAS,SAAS,UAAU;AAClC,UAAQ,IAAI;GACV,QAAQ;GACR,MAAM;GACN,SAAS,eAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE;IAAQ;IAAS;GAC9B,CAAC;AACF,KAAG,KACD,KAAK,UACH,gBAAgB,SAAS,MAAM,SAAS,SAAS,MAAM,MAAM,SAAS,MAAM,KAAK,CAClF,CACF;AACD;;AAIF,KAAI,+BAA+B,SAAS,EAAE;EAC5C,MAAM,eAAe,QAAQ,IAAI;GAC/B,QAAQ;GACR,MAAM;GACN,SAAS,eAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE,QAAQ;IAAK;IAAS;GACnC,CAAC;EAEF,MAAM,SAAS,sCACb,SAAS,WAAW,IACpB,SAAS,aAAa,EAAE,EACxB,cAAc,OACd,WACA,yBACE,SAAS,WACT,cAAc,OACd,iBACA,SAAS,OACV,EACD,SAAS,aACT,iBAAiB,SAAS,EAC1B,SAAS,QACT,wBACA,+BACD;EAED,MAAM,eAAe,yBAAyB,QAAQ;AAUtD,MAAI,CATc,MAAM,WACtB,IACA,QACA,SACA,cAAc,QACd,cAAc,MACd,QAAQ,iBACR,QAAQ,eAAe,SAAS,YACjC,EACe;AACd,MAAG,SAAS;AACZ,gBAAa,SAAS,cAAc;AACpC,gBAAa,SAAS,kBAAkB,cAAc,QAAQ;;AAEhE,gBAAc,SAAS;AACvB;;AAIF,KAAI,eAAe,SAAS,EAAE;EAC5B,MAAM,eAAe,QAAQ,IAAI;GAC/B,QAAQ;GACR,MAAM;GACN,SAAS,eAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE,QAAQ;IAAK;IAAS;GACnC,CAAC;EAEF,MAAM,SAAS,sBACb,SAAS,SACT,cAAc,OACd,WACA,yBACE,SAAS,WACT,cAAc,OACd,iBACA,SAAS,OACV,EACD,SAAS,aACT,iBAAiB,SAAS,EAC1B,wBACA,+BACD;EACD,MAAM,eAAe,yBAAyB,QAAQ;AAUtD,MAAI,CATc,MAAM,WACtB,IACA,QACA,SACA,cAAc,QACd,cAAc,MACd,QAAQ,iBACR,QAAQ,eAAe,SAAS,YACjC,EACe;AACd,MAAG,SAAS;AACZ,gBAAa,SAAS,cAAc;AACpC,gBAAa,SAAS,kBAAkB,cAAc,QAAQ;;AAEhE,gBAAc,SAAS;AACvB;;AAIF,KAAI,mBAAmB,SAAS,EAAE;EAChC,MAAM,eAAe,QAAQ,IAAI;GAC/B,QAAQ;GACR,MAAM;GACN,SAAS,eAAe,SAAS,kBAAkB,EAAE,CAAC;GACtD,MAAM;GACN,UAAU;IAAE,QAAQ;IAAK;IAAS;GACnC,CAAC;EACF,MAAM,SAAS,0BACb,SAAS,WACT,cAAc,OACd,WAIA,yBACE,SAAS,WACT,cAAc,OACd,iBACA,SAAS,OACV,EACD,SAAS,aACT,iBAAiB,SAAS,EAC1B,wBACA,+BACD;EACD,MAAM,eAAe,yBAAyB,QAAQ;AAUtD,MAAI,CATc,MAAM,WACtB,IACA,QACA,SACA,cAAc,QACd,cAAc,MACd,QAAQ,iBACR,QAAQ,eAAe,SAAS,YACjC,EACe;AACd,MAAG,SAAS;AACZ,gBAAa,SAAS,cAAc;AACpC,gBAAa,SAAS,kBAAkB,cAAc,QAAQ;;AAEhE,gBAAc,SAAS;AACvB;;AAIF,SAAQ,IAAI;EACV,QAAQ;EACR,MAAM;EACN,SAAS,eAAe,SAAS,kBAAkB,EAAE,CAAC;EACtD,MAAM;EACN,UAAU;GAAE,QAAQ;GAAK;GAAS;EACnC,CAAC;AACF,IAAG,KACD,KAAK,UACH,gBAAgB,iDAAiD,eAAe,CACjF,CACF;;AAGH,eAAe,WACb,IACA,QACA,SACA,QACA,aACA,iBACA,aACkB;CAClB,IAAI,aAAa;AACjB,MAAK,MAAM,SAAS,QAAQ;AAC1B,MAAI,GAAG,SAAU,QAAO;EACxB,MAAM,aAAa,eAAe,YAAY,QAAW,SAAS,iBAAiB,YAAY;AAC/F,MAAI,aAAa,EAAG,OAAM,MAAM,YAAY,OAAO;AACnD,MAAI,QAAQ,QAAS,QAAO;AAC5B,MAAI,GAAG,SAAU,QAAO;AACxB,KAAG,KAAK,KAAK,UAAU,MAAM,CAAC;AAC9B;AACA,iBAAe;AACf,MAAI,QAAQ,QAAS,QAAO;;AAE9B,QAAO"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@copilotkit/aimock",
-  "version": "1.37.3",
+  "version": "1.38.0",
   "description": "Mock infrastructure for AI application testing — LLM APIs, image generation, image editing, text-to-speech, transcription, audio translation, audio generation, video generation, embeddings, MCP tools, A2A agents, AG-UI event streams, vector databases, search, rerank, and moderation. One package, one port, zero dependencies.",
   "license": "MIT",
   "keywords": [

package/skills/write-fixtures/SKILL.md

@@ -7,7 +7,7 @@ description: Use when writing test fixtures for @copilotkit/aimock — mock LLM
 
 ## What aimock Is
 
-aimock is a zero-dependency mock infrastructure for AI apps. Fixture-driven. Multi-provider (OpenAI, Anthropic, Gemini, Gemini Interactions, AWS Bedrock, Azure OpenAI, Vertex AI, Ollama, Cohere). Multimedia endpoints (image generation, text-to-speech, audio transcription, video generation). MCP, A2A, AG-UI, and vector DB mocking. Runs a real HTTP server on a real port — works across processes, unlike MSW-style interceptors. WebSocket support for OpenAI Responses/Realtime and Gemini Live APIs. Record-and-replay for all endpoints including multimedia. Chaos testing and Prometheus metrics.
+aimock is a zero-dependency mock infrastructure for AI apps. Fixture-driven. Multi-provider (OpenAI, Anthropic, Gemini, Gemini Interactions, AWS Bedrock, Azure OpenAI, Vertex AI, Ollama, Cohere, OpenRouter). Multimedia endpoints (image generation, text-to-speech, audio transcription, video generation). MCP, A2A, AG-UI, and vector DB mocking. Runs a real HTTP server on a real port — works across processes, unlike MSW-style interceptors. WebSocket support for OpenAI Responses/Realtime and Gemini Live APIs. Record-and-replay for all endpoints including multimedia. Chaos testing and Prometheus metrics.
 
 ## Core Mental Model
 
@@ -19,26 +19,26 @@ aimock is a zero-dependency mock infrastructure for AI apps. Fixture-driven. Mul
 
 ## Match Field Reference
 
-| Field                | Type                                      | Matches Against                                                                                                                                                                                                                                                                                                  |
-| -------------------- | ----------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| `userMessage`        | `string`                                  | Substring of last `role: "user"` message text                                                                                                                                                                                                                                                                    |
-| `userMessage`        | `RegExp`                                  | Pattern test on last `role: "user"` message text                                                                                                                                                                                                                                                                 |
-| `systemMessage`      | `string`                                  | Substring of the concatenated text of every `role: "system"` message in the request. Use to gate a fixture on host-supplied context (persona, agent-context entries) so changes to that context cause the fixture to fall through instead of returning a stale baked response                                    |
-| `systemMessage`      | `string[]`                                | Array of substrings — ALL must be present in the joined system text (AND semantics). Use when the gate must combine multiple non-adjacent tokens whose serialisation order isn't stable                                                                                                                          |
-| `systemMessage`      | `RegExp`                                  | Pattern test on the concatenated system-message text                                                                                                                                                                                                                                                             |
-| `inputText`          | `string`                                  | Substring of embedding input text (concatenated if multiple inputs)                                                                                                                                                                                                                                              |
-| `inputText`          | `RegExp`                                  | Pattern test on embedding input text                                                                                                                                                                                                                                                                             |
-| `toolName`           | `string`                                  | Exact match on any tool in request's `tools[]` array (by `function.name`)                                                                                                                                                                                                                                        |
-| `toolCallId`         | `string`                                  | Exact match on `tool_call_id` of last `role: "tool"` message                                                                                                                                                                                                                                                     |
-| `toolResultContains` | `string`                                  | Substring of the last tool message's text content, gated on that message being the request's LAST message (same rule as `toolCallId`). Discriminates resume paths that share a `tool_call_id` and differ only inside the tool-result payload (e.g. approve `{"chosen_time": …}` vs cancel `{"cancelled": true}`) |
-| `model`              | `string`                                  | Exact match on `req.model`                                                                                                                                                                                                                                                                                       |
-| `model`              | `RegExp`                                  | Pattern test on `req.model`                                                                                                                                                                                                                                                                                      |
-| `responseFormat`     | `string`                                  | Exact match on `req.response_format.type` (`"json_object"`, `"json_schema"`)                                                                                                                                                                                                                                     |
-| `sequenceIndex`      | `number`                                  | Matches only when this fixture's match count equals the given index (0-based)                                                                                                                                                                                                                                    |
-| `turnIndex`          | `number`                                  | Stateless conversation-depth matching. Counts `role: "assistant"` messages in the request; matches when that count equals the value. `turnIndex: 0` = first turn (no prior assistant messages). Use instead of `sequenceIndex` for shared/deployed instances where stateful counters break under concurrency     |
-| `hasToolResult`      | `boolean`                                 | Stateless tool-message presence matching. `true` matches when any `role: "tool"` message exists in the request; `false` matches when none exist. Provider-consistent across all aimock handlers (OpenAI, Claude, Gemini, Bedrock, Ollama, Cohere)                                                                |
-| `endpoint`           | `string`                                  | Restrict to endpoint type: `"chat"`, `"image"`, `"speech"`, `"transcription"`, `"video"`, `"embedding"`                                                                                                                                                                                                          |
-| `predicate`          | `(req: ChatCompletionRequest) => boolean` | Custom function — full access to request                                                                                                                                                                                                                                                                         |
+| Field                | Type                                      | Matches Against                                                                                                                                                                                                                                                                                                                                                                                                    |
+| -------------------- | ----------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
+| `userMessage`        | `string`                                  | Substring of last `role: "user"` message text                                                                                                                                                                                                                                                                                                                                                                      |
+| `userMessage`        | `RegExp`                                  | Pattern test on last `role: "user"` message text                                                                                                                                                                                                                                                                                                                                                                   |
+| `systemMessage`      | `string`                                  | Substring of the concatenated text of every `role: "system"` message in the request. Use to gate a fixture on host-supplied context (persona, agent-context entries) so changes to that context cause the fixture to fall through instead of returning a stale baked response                                                                                                                                      |
+| `systemMessage`      | `string[]`                                | Array of substrings — ALL must be present in the joined system text (AND semantics). Use when the gate must combine multiple non-adjacent tokens whose serialisation order isn't stable                                                                                                                                                                                                                            |
+| `systemMessage`      | `RegExp`                                  | Pattern test on the concatenated system-message text                                                                                                                                                                                                                                                                                                                                                               |
+| `inputText`          | `string`                                  | Substring of embedding input text (concatenated if multiple inputs)                                                                                                                                                                                                                                                                                                                                                |
+| `inputText`          | `RegExp`                                  | Pattern test on embedding input text                                                                                                                                                                                                                                                                                                                                                                               |
+| `toolName`           | `string`                                  | Exact match on any tool in request's `tools[]` array (by `function.name`)                                                                                                                                                                                                                                                                                                                                          |
+| `toolCallId`         | `string`                                  | Exact match on `tool_call_id` of last `role: "tool"` message                                                                                                                                                                                                                                                                                                                                                       |
+| `toolResultContains` | `string`                                  | Substring of the last tool message's text content, gated on that message being the request's LAST message (same rule as `toolCallId`). Discriminates resume paths that share a `tool_call_id` and differ only inside the tool-result payload (e.g. approve `{"chosen_time": …}` vs cancel `{"cancelled": true}`)                                                                                                   |
+| `model`              | `string`                                  | Exact match on `req.model`                                                                                                                                                                                                                                                                                                                                                                                         |
+| `model`              | `RegExp`                                  | Pattern test on `req.model`                                                                                                                                                                                                                                                                                                                                                                                        |
+| `responseFormat`     | `string`                                  | Exact match on `req.response_format.type` (`"json_object"`, `"json_schema"`)                                                                                                                                                                                                                                                                                                                                       |
+| `sequenceIndex`      | `number`                                  | Matches only when this fixture's match count equals the given index (0-based)                                                                                                                                                                                                                                                                                                                                      |
+| `turnIndex`          | `number`                                  | Stateless conversation-depth matching. Counts `role: "assistant"` messages in the request; matches when that count equals the value. `turnIndex: 0` = first turn (no prior assistant messages). Use instead of `sequenceIndex` for shared/deployed instances where stateful counters break under concurrency                                                                                                       |
+| `hasToolResult`      | `boolean`                                 | Stateless tool-message presence matching, scoped to the CURRENT turn (messages after the last `role: "user"` message). `true` matches when a `role: "tool"` message appears after the last user message; `false` matches when none does. (If the request has no user message, the whole conversation is scanned.) Provider-consistent across all aimock handlers (OpenAI, Claude, Gemini, Bedrock, Ollama, Cohere) |
+| `endpoint`           | `string`                                  | Restrict to endpoint type: `"chat"`, `"image"`, `"speech"`, `"transcription"`, `"video"`, `"embedding"`                                                                                                                                                                                                                                                                                                            |
+| `predicate`          | `(req: ChatCompletionRequest) => boolean` | Custom function — full access to request                                                                                                                                                                                                                                                                                                                                                                           |
 
 **AND logic**: all specified fields must match. Empty match `{}` = catch-all.
 
@@ -46,13 +46,13 @@ Multi-part content (e.g., `[{type: "text", text: "hello"}]`) is automatically ex
 
 ### When to Use Each Multi-turn Matching Approach
 
-| Approach             | Stateless? | Best For                                                                                                  |
-| -------------------- | ---------- | --------------------------------------------------------------------------------------------------------- |
-| `turnIndex`          | Yes        | Shared/deployed instances; matches on conversation depth (count of assistant messages in request)         |
-| `hasToolResult`      | Yes        | Simplest option for 2-step tool flows — boolean: are there tool results in the request?                   |
-| `sequenceIndex`      | No         | Single-client unit tests with repeated identical requests (server-side counter, breaks under concurrency) |
-| `toolCallId`         | Yes        | Matching specific tool result IDs in the conversation history                                             |
-| `toolResultContains` | Yes        | Same tool call id, different outcomes — match on the tool-result payload (approve vs cancel legs)         |
+| Approach             | Stateless? | Best For                                                                                                                  |
+| -------------------- | ---------- | ------------------------------------------------------------------------------------------------------------------------- |
+| `turnIndex`          | Yes        | Shared/deployed instances; matches on conversation depth (count of assistant messages in request)                         |
+| `hasToolResult`      | Yes        | Simplest option for 2-step tool flows — boolean: does the current turn (after the last user message) carry a tool result? |
+| `sequenceIndex`      | No         | Single-client unit tests with repeated identical requests (server-side counter, breaks under concurrency)                 |
+| `toolCallId`         | Yes        | Matching specific tool result IDs in the conversation history                                                             |
+| `toolResultContains` | Yes        | Same tool call id, different outcomes — match on the tool-result payload (approve vs cancel legs)                         |
 
 **Prefer stateless approaches** (`turnIndex`, `hasToolResult`, `toolResultContains`) for shared aimock instances (deployed via Docker, used by multiple test runners). Use `sequenceIndex` only in isolated single-client unit tests where the counter won't be corrupted by concurrent requests.
 
@@ -420,6 +420,8 @@ All providers share the same fixture pool — write fixtures once, they work for
 | `POST /api/generate`                                                                     | Ollama        | HTTP      |
 | `GET /api/tags`                                                                          | Ollama        | HTTP      |
 | `POST /v2/chat`                                                                          | Cohere        | HTTP      |
+| `POST /api/v1/chat/completions`                                                          | OpenRouter    | HTTP      |
+| `GET /api/v1/models` · `/api/v1/key` · `/api/v1/credits`                                 | OpenRouter    | HTTP      |
 | `GET /metrics`                                                                           | —             | HTTP      |
 | `GET /v1/models`                                                                         | OpenAI-compat | HTTP      |
 | `WS /v1/responses`                                                                       | OpenAI        | WebSocket |
@@ -436,15 +438,19 @@ All providers share the same fixture pool — write fixtures once, they work for
 
 Fixture responses can include optional override fields to control auto-generated envelope values. These are merged into the provider-specific response format (OpenAI, Claude, Gemini, Responses API).
 
-| Field               | Type   | Default                   | Description                                                                                                                                                                                                                                                                |
-| ------------------- | ------ | ------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| `id`                | string | auto-generated            | Override response ID (e.g., `chatcmpl-custom`)                                                                                                                                                                                                                             |
-| `created`           | number | `Date.now()/1000`         | Override Unix timestamp                                                                                                                                                                                                                                                    |
-| `model`             | string | echoes request            | Override model name in response                                                                                                                                                                                                                                            |
-| `usage`             | object | zeroed                    | Override token counts: `{ prompt_tokens, completion_tokens, total_tokens }`. OpenAI Chat includes usage in response body; Responses API uses `response.usage`. When omitted, auto-computed from content length                                                             |
-| `finishReason`      | string | `"stop"` / `"tool_calls"` | Override finish reason. Mappings: `stop` -> `end_turn` (Claude), `STOP` (Gemini); `tool_calls` -> `tool_use` (Claude), `FUNCTION_CALL` (Gemini); `length` -> `max_tokens` (Claude), `MAX_TOKENS` (Gemini); `content_filter` -> `SAFETY` (Gemini), `failed` (Responses API) |
-| `role`              | string | `"assistant"`             | Override message role                                                                                                                                                                                                                                                      |
-| `systemFingerprint` | string | (omitted)                 | Add `system_fingerprint` to response                                                                                                                                                                                                                                       |
+| Field                | Type   | Default                   | Description                                                                                                                                                                                                                                                                |
+| -------------------- | ------ | ------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| `id`                 | string | auto-generated            | Override response ID (e.g., `chatcmpl-custom`)                                                                                                                                                                                                                             |
+| `created`            | number | `Date.now()/1000`         | Override Unix timestamp                                                                                                                                                                                                                                                    |
+| `model`              | string | echoes request            | Override model name in response                                                                                                                                                                                                                                            |
+| `usage`              | object | zeroed                    | Override token counts: `{ prompt_tokens, completion_tokens, total_tokens }`. OpenAI Chat includes usage in response body; Responses API uses `response.usage`. When omitted, auto-computed from content length                                                             |
+| `finishReason`       | string | `"stop"` / `"tool_calls"` | Override finish reason. Mappings: `stop` -> `end_turn` (Claude), `STOP` (Gemini); `tool_calls` -> `tool_use` (Claude), `FUNCTION_CALL` (Gemini); `length` -> `max_tokens` (Claude), `MAX_TOKENS` (Gemini); `content_filter` -> `SAFETY` (Gemini), `failed` (Responses API) |
+| `role`               | string | `"assistant"`             | Override message role                                                                                                                                                                                                                                                      |
+| `systemFingerprint`  | string | (omitted)                 | Add `system_fingerprint` to response                                                                                                                                                                                                                                       |
+| `provider`           | string | slug author               | OpenRouter only: top-level serving-provider display name (default = the winning model slug's author). Override to assert who served the request                                                                                                                            |
+| `nativeFinishReason` | string | mirrors `finishReason`    | OpenRouter only: the raw upstream `native_finish_reason` alongside the normalized `finish_reason`                                                                                                                                                                          |
+| `usage.cost`         | number | (omitted)                 | OpenRouter only: per-request `usage.cost` (scriptable — powers budget-guard tests). When set, `usage.cost_details` is emitted too. Never fabricated when omitted                                                                                                           |
+| `usage.is_byok`      | bool   | (omitted)                 | OpenRouter only: emit `usage.is_byok`. Also `usage.prompt_tokens_details`, `usage.completion_tokens_details` — emitted only when set                                                                                                                                       |
 
 ### Example
 
@@ -473,17 +479,45 @@ mock.onMessage("hello", {
 
 These fields map correctly across all provider formats — for example, `finishReason: "stop"` becomes `finish_reason: "stop"` in OpenAI, `stop_reason: "end_turn"` in Claude, and `finishReason: "STOP"` in Gemini.
 
+## OpenRouter (chat / router)
+
+A request whose path starts with `/api/v1/` (point the OpenAI SDK at a `baseURL` ending `/api/v1`) is shaped as OpenRouter: `gen-` id, top-level `provider`, per-choice `native_finish_reason`, always-present `system_fingerprint`/`service_tier` (null by default), an always-present `message.reasoning` (null unless a fixture supplies reasoning and the model is reasoning-capable), and a rich `usage`. Requests on the plain `/v1/...` base are untouched OpenAI. Same fixture pool — the fields above are the only additions.
+
+- **Scriptable cost / provider / finish reason**: set `provider`, `nativeFinishReason`, and `usage.cost` on the response (see the overrides table). `cost`/`cost_details` are emitted only when a fixture supplies `cost` — aimock never fabricates a cost.
+- **`models[]` fallback (router failover)**: when the request body carries `models: [m1, m2, ...]`, aimock walks `[model, ...models]` in order and serves the first fixture that returns a NON-error response. A `429`/`503` error fixture on a candidate simulates a RUNTIME provider failure and falls through to the next candidate; the winning slug is echoed back as the top-level `model` (assert failover via `response.model`). Model the primary's "failure" as a 429/503 — an unknown/invalid model is just a fixture miss (aimock does not replicate OpenRouter's up-front invalid-model 400).
+- **Terminal (non-failover) error class — `fallthrough: false`**: real OpenRouter fails over inconsistently by error class (a `403 budget-exceeded` / generic "provider returned error" is served as terminal and does NOT advance to the next candidate, while 429/503 usually do). Set `fallthrough: false` on an error fixture to make it terminal: the fallback loop stops and serves that error even when a good candidate follows. Absent / `true` keeps the default fall-through. Composes with `provider.allow_fallbacks`: fall-through happens only when both allow it (if either says stop, the error is terminal). Use it to reproduce the exact provider error a dev's app must handle itself. `{ error: { message: "budget exceeded" }, status: 403, fallthrough: false }`
+- **Keepalive**: set the fixture option `openRouterProcessing: true` to emit one `: OPENROUTER PROCESSING` SSE comment before the first data frame (opt-in, default off).
+- **Error envelope**: OpenRouter errors are `{ error: { message, code } }` (numeric `code` == HTTP status), with optional free-form `metadata`.
+
+```typescript
+// primary is a runtime 429, fallback answers
+mock.on(
+  { model: "openai/gpt-4o", userMessage: "route" },
+  { error: { message: "rate limited" }, status: 429 },
+);
+mock.on(
+  { model: "anthropic/claude-3.5-sonnet", userMessage: "route" },
+  {
+    content: "served by the fallback",
+    provider: "Anthropic",
+    usage: { cost: 0.0021 },
+  },
+);
+// POST /api/v1/chat/completions { model: "openai/gpt-4o", models: [...], ... }
+//   → response.model === "anthropic/claude-3.5-sonnet"
+```
+
 ## Provider Support Matrix
 
-| Feature              | OpenAI Chat | OpenAI Responses | Claude | Gemini | Gemini Int. | Bedrock | Azure | Ollama | Cohere |
-| -------------------- | ----------- | ---------------- | ------ | ------ | ----------- | ------- | ----- | ------ | ------ |
-| Text                 | Yes         | Yes              | Yes    | Yes    | Yes         | Yes     | Yes   | Yes    | Yes    |
-| Tool Calls           | Yes         | Yes              | Yes    | Yes    | Yes         | Yes     | Yes   | Yes    | Yes    |
-| Content + Tool Calls | Yes         | Yes              | Yes    | Yes    | Yes         | Yes     | Yes   | Yes    | Yes    |
-| Streaming            | SSE         | SSE              | SSE    | SSE    | SSE         | Binary  | SSE   | NDJSON | SSE    |
-| Reasoning            | Yes         | Yes              | Yes    | Yes    | --          | Yes     | Yes   | --     | --     |
-| Web Searches         | --          | Yes              | --     | --     | --          | --      | --    | --     | --     |
-| Response Overrides   | Yes         | Yes              | Yes    | Yes    | Yes         | --      | Yes   | --     | --     |
+| Feature              | OpenAI Chat | OpenAI Responses | Claude | Gemini | Gemini Int. | Bedrock | Azure | Ollama | Cohere | OpenRouter |
+| -------------------- | ----------- | ---------------- | ------ | ------ | ----------- | ------- | ----- | ------ | ------ | ---------- |
+| Text                 | Yes         | Yes              | Yes    | Yes    | Yes         | Yes     | Yes   | Yes    | Yes    | Yes        |
+| Tool Calls           | Yes         | Yes              | Yes    | Yes    | Yes         | Yes     | Yes   | Yes    | Yes    | Yes        |
+| Content + Tool Calls | Yes         | Yes              | Yes    | Yes    | Yes         | Yes     | Yes   | Yes    | Yes    | Yes        |
+| Streaming            | SSE         | SSE              | SSE    | SSE    | SSE         | Binary  | SSE   | NDJSON | SSE    | SSE        |
+| Reasoning            | Yes         | Yes              | Yes    | Yes    | --          | Yes     | Yes   | --     | --     | Yes        |
+| Web Searches         | --          | Yes              | --     | --     | --          | --      | --    | --     | --     | --         |
+| Response Overrides   | Yes         | Yes              | Yes    | Yes    | Yes         | --      | Yes   | --     | --     | Yes        |
 
 ## Critical Gotchas
 
@@ -495,7 +529,7 @@ These fields map correctly across all provider formats — for example, `finishR
 
 4. **`streamingProfile` takes precedence over `latency`** — when both are set on a fixture, `streamingProfile` controls timing. Use one or the other.
 
-5. **Tool result messages don't change the user message** — after a tool call, the client sends the same conversation + tool result. Matching on `userMessage` will hit the SAME fixture again → infinite loop. Always use `predicate` checking `role === "tool"` for tool results.
+5. **Tool result messages don't change the user message** — after a tool call, the client sends the same conversation + tool result. Matching on `userMessage` will hit the SAME fixture again → infinite loop. Always use `predicate` checking `role === "tool"` for tool results. Note: a whole-conversation `role === "tool"` check (e.g. `req.messages.some((m) => m.role === "tool")`) diverges from `hasToolResult`'s current-turn scoping in multi-turn flows — the built-in `hasToolResult` matcher only looks after the last user message, so a later turn whose history carries an earlier tool result still reads `false`.
 
 6. **`clearFixtures()` preserves the array reference** — uses `.length = 0`, not reassignment. The running server reads the same array object.