> ## Documentation Index > Fetch the complete documentation index at: https://docs.signalmd.dev/llms.txt > Use this file to discover all available pages before exploring further. # Get an attempt > One simulated call: its transcript, the verdict on each criterion with who decided it, what SignalMD's own checks found, caller drift, and what played and scored it. ## OpenAPI ````yaml /api-reference/openapi.json get /v1/agents/{agentId}/attempts/{attemptId} openapi: 3.1.0 info: title: SignalMD API version: 1.0.0 description: >- SignalMD reads a voice agent's production calls and says what to fix, then tests whether the fix worked. This API exposes the same data the web app shows: agents, calls and their traces, signals, issues, versions and their comparisons, tests, and test runs. Authenticate with `Authorization: Bearer sig_...` (an API key from the API keys page). Read keys may call every `read` operation; write keys may also call `write` operations, within their monthly budget. Each operation's scope is in `x-signal-scope`. Errors share one envelope (`Error`), lists share one cursor scheme (`limit`, `cursor`, `hasMore`, `nextCursor`), and every response carries `Request-Id` and `RateLimit-*` headers. servers: - url: https://api.signalmd.dev description: Production - url: http://localhost:43124 description: Local development security: - bearerAuth: [] tags: - name: Account description: Who the token belongs to and what it may do. - name: Agents description: Voice agents connected to the workspace. - name: Calls description: Production calls, their traces and per-call signals. - name: Signals description: What SignalMD noticed across calls. - name: Issues description: 'Problems worth fixing: where, why, and what to change.' - name: Versions description: Agent versions and how each one did. - name: Comparisons description: Stored comparisons between two versions. - name: Tests description: Tests, test sets, personas and coverage. - name: Runs description: Test runs and the attempts inside them. paths: /v1/agents/{agentId}/attempts/{attemptId}: get: tags: - Runs summary: Get an attempt description: >- One simulated call: its transcript, the verdict on each criterion with who decided it, what SignalMD's own checks found, caller drift, and what played and scored it. operationId: getAttempt parameters: - schema: type: string format: uuid pattern: >- ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$ example: 3f1c2d4e-5a6b-4c7d-8e9f-0a1b2c3d4e5f description: The agent's id. required: true description: The agent's id. name: agentId in: path - schema: type: string format: uuid pattern: >- ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$ example: 3f1c2d4e-5a6b-4c7d-8e9f-0a1b2c3d4e5f description: The attempt's id. required: true description: The attempt's id. name: attemptId in: path responses: '200': description: The attempt. content: application/json: schema: $ref: '#/components/schemas/Attempt' '400': description: 'Invalid request: a parameter or the body failed validation.' content: application/json: schema: $ref: '#/components/schemas/Error' '401': description: Missing, invalid or revoked credentials. content: application/json: schema: $ref: '#/components/schemas/Error' '403': description: Not permitted for this principal. content: application/json: schema: $ref: '#/components/schemas/Error' '404': description: >- The agent, or a resource under it, does not exist or is not visible to this key. content: application/json: schema: $ref: '#/components/schemas/Error' '429': description: Rate limited. See Retry-After. content: application/json: schema: $ref: '#/components/schemas/Error' '500': description: Server error. The message is generic; quote the request id. content: application/json: schema: $ref: '#/components/schemas/Error' x-codeSamples: - lang: bash label: curl source: >- curl "https://api.signalmd.dev/v1/agents/6f1d2c3b-4a5e-4f60-8b7a-9c0d1e2f3a4b/attempts/f9a0b1c2-d3e4-4f5a-8b6c-7d8e9f0a1b2c" \ -H "Authorization: Bearer $SIGNAL_API_KEY" components: schemas: Attempt: allOf: - $ref: '#/components/schemas/AttemptSummary' - type: object properties: transcript: description: >- The conversation as the voice platform returned it, tool calls included. Its shape is the platform's, not normalised. verdicts: type: array items: $ref: '#/components/schemas/CriterionVerdict' findings: type: array items: type: object additionalProperties: {} description: >- Problems SignalMD's own checks found on the transcript (code, title, severity, detail), whatever the criteria say. drift: type: - object - 'null' additionalProperties: {} description: >- Whether the simulated caller stayed the source caller (drifted, reasons, notes). Null: no source to compare with. scoreHistory: type: array items: type: object additionalProperties: {} description: >- Earlier scores of this attempt, oldest first, each with the epoch it was scored in. watches: type: array items: $ref: '#/components/schemas/AttemptWatch' description: >- Live watches of the agent on this transcript, reported beside the verdict. A watch never changes the verdict. required: - verdicts - findings - drift - scoreHistory - watches Error: type: object properties: error: type: object properties: type: $ref: '#/components/schemas/ErrorType' code: type: string example: agent_not_found description: The specific reason; stable within v1. message: type: string example: Agent not found. param: type: - string - 'null' description: The request field at fault, for a validation error. requestId: type: - string - 'null' example: req_8c1f0e2b4a6d9e3f7a5c1b0d quota: type: object properties: metric: type: string example: credits description: >- credits when the workspace's monthly credits are used up, or agents when the plan's agent limit is reached. activity: type: string enum: - observation - testing - chat example: testing description: >- With metric credits: what the request would have spent them on. observation is analysing calls, testing is test runs, chat is the copilot. requested: type: integer example: 18 description: >- With metric credits: the credits the request needed to start. limit: type: integer description: 'The plan''s limit: the month''s credits, or its agents.' used: type: integer description: How much of the limit is used this month. plan: type: string example: free_org description: The workspace's plan slug. upgradePath: type: string example: /billing description: Where in the web app the plan is changed. required: - metric - limit - used - plan - upgradePath description: >- Present when code is quota_exceeded: the plan limit the request would pass. proposalId: type: string format: uuid description: >- Present when code is criteria_locked: the pending criteria proposal the change was filed as. A person approves or rejects it. required: - type - code - message - param - requestId required: - error example: error: type: not_found_error code: agent_not_found message: Agent not found. param: null requestId: req_8c1f0e2b4a6d9e3f7a5c1b0d AttemptSummary: type: object properties: object: type: string enum: - attempt id: type: string format: uuid pattern: >- ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$ example: 3f1c2d4e-5a6b-4c7d-8e9f-0a1b2c3d4e5f runId: type: - string - 'null' format: uuid pattern: >- ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$ example: 3f1c2d4e-5a6b-4c7d-8e9f-0a1b2c3d4e5f description: The run it belongs to. testId: type: string format: uuid pattern: >- ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$ example: 3f1c2d4e-5a6b-4c7d-8e9f-0a1b2c3d4e5f testRevision: type: - integer - 'null' description: The test revision it ran. Null on attempts stored before revisions. persona: type: string description: The persona slug played, or @source for the test's own caller. attempt: type: integer description: >- 1 for the first try; 2 for the rerun of a failure; 2 and 3 for the repeats of a cell that runs three times. version: type: - integer - 'null' description: The agent version played. result: type: - string - 'null' enum: - passed - failed - error - drift - null description: >- This attempt's own result. The cell's combined result, including flaky, is in the run's result document. reason: type: - string - 'null' description: Why, in one line. noFixture: type: array items: type: string description: Tools that answered with the fixed 'not available' error. providerGrader: type: object properties: status: type: - string - 'null' explanation: type: - string - 'null' required: - status - explanation description: The voice platform's own grader result. models: type: object properties: simulator: type: - string - 'null' description: What played the caller. judge: type: - string - 'null' description: How the criteria were scored. judgePromptHash: type: - string - 'null' description: >- Identifier of the review setup used, for comparing runs. Runs with different values were scored differently. scorerVersion: type: - string - 'null' required: - simulator - judge - judgePromptHash - scorerVersion description: What played and scored it. cost: type: object properties: retellUsd: type: - number - 'null' signalUsd: type: - number - 'null' required: - retellUsd - signalUsd description: 'Estimated cost of the attempt: Retell''s share and SignalMD''s.' rescoredAt: type: - string - 'null' format: date-time example: '2026-10-02T14:03:07.120Z' description: Set when it was scored again from its transcript. startedAt: type: string format: date-time example: '2026-10-02T14:03:07.120Z' finishedAt: type: - string - 'null' format: date-time example: '2026-10-02T14:03:07.120Z' required: - object - id - runId - testId - testRevision - persona - attempt - version - result - reason - noFixture - providerGrader - models - cost - rescoredAt - startedAt - finishedAt example: object: attempt id: f9a0b1c2-d3e4-4f5a-8b6c-7d8e9f0a1b2c runId: d1e2f3a4-b5c6-4d7e-8f9a-0b1c2d3e4f5a testId: 3f1c2d4e-5a6b-4c7d-8e9f-0a1b2c3d4e5f testRevision: 2 persona: '@source' attempt: 1 version: 8 result: passed reason: null noFixture: [] providerGrader: status: null explanation: null models: simulator: caller-setup-a judge: review-setup-a judgePromptHash: 9f2c1a7e scorerVersion: '2026-09-18' cost: retellUsd: 0.15 signalUsd: 0.02 rescoredAt: null startedAt: '2026-10-02T11:01:15.000Z' finishedAt: '2026-10-02T11:02:40.000Z' CriterionVerdict: type: object properties: id: type: string criterionId: type: string text: type: string decidedBy: type: string enum: - detector - judge - unmeasured description: >- How the criterion was decided: `detector` by a check on the transcript and tool calls, `judge` by reading the conversation, `unmeasured` not at all, because a text run cannot measure it. met: type: - boolean - 'null' description: >- Null: not decided (not measurable in text, or the answer was unclear). detail: type: string judge: type: object properties: model: type: - string - 'null' description: The review setup that read the conversation for this criterion. answers: type: array items: {} description: Each answer given when the criterion was scored. evidenceTurn: type: integer quote: type: string evidenceKind: type: string required: - model - answers description: >- Set when the criterion was scored by reading the conversation, or could not be. required: - id - criterionId - text - decidedBy - met - detail AttemptWatch: type: object properties: watchId: type: string description: The watch's id. name: type: string description: The watch's name. status: type: string description: >- What the watch found on this transcript: met, violated, not in scope, not checked, checking, unclear, or flagged. "Not checked: no model" means SignalMD could not check this call. quote: type: - string - 'null' description: >- Why the watch was broken, quoted from the call, when status is violated; otherwise null. evidenceTurn: type: - integer - 'null' description: The turn the quote comes from, when it comes from one turn. required: - watchId - name - status - quote ErrorType: type: string enum: - invalid_request_error - authentication_error - budget_error - permission_error - not_found_error - conflict_error - rate_limit_error - api_error securitySchemes: bearerAuth: type: http scheme: bearer description: An API key (`sig_...`), or a session token from the web app. ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.