> ## Documentation Index > Fetch the complete documentation index at: https://docs.upliftai.org/llms.txt > Use this file to discover all available pages before exploring further. # Poll a build's progress > Works for refine runs too — poll whichever `buildId` you got back. Runs take minutes, so poll every few seconds. `done` only means the pipeline stopped; whether the grader was satisfied is `result.passed`. Live detail (`phase`, `round`, transcript counts) is kept for an hour after the run last moved; after that the poll still returns the outcome — `result` on success, `error` on failure. ## OpenAPI ````yaml /voice-agents/api-reference/openapi.yaml get /realtime-assistants/builds/status/{buildId} openapi: 3.1.0 info: title: UpliftAI Voice Agents API version: '1.0' description: >- Build voice assistants, start conversations on the web or the phone, run calling campaigns, and receive outcomes by webhook. All endpoints take a bearer API key unless noted otherwise. servers: - url: https://api.upliftai.org/v1 security: - apiKey: [] tags: - name: Assistants - name: Authoring an assistant - name: Starting a conversation - name: Sessions & call records - name: Campaigns - name: Campaign outcomes - name: Contacts - name: Google Sheet sources - name: Integrations & delivery health paths: /realtime-assistants/builds/status/{buildId}: get: tags: - Authoring an assistant summary: Poll a build's progress description: >- Works for refine runs too — poll whichever `buildId` you got back. Runs take minutes, so poll every few seconds. `done` only means the pipeline stopped; whether the grader was satisfied is `result.passed`. Live detail (`phase`, `round`, transcript counts) is kept for an hour after the run last moved; after that the poll still returns the outcome — `result` on success, `error` on failure. operationId: getAssistantBuildStatus parameters: - name: buildId in: path required: true description: The `buildId` the build or refine returned. schema: type: string format: uuid example: 425d3b75-xxxx-xxxx-xxxx-xxxxxxxxxxxx responses: '200': description: The run's latest frame. content: application/json: schema: $ref: '#/components/schemas/BuildProgress' examples: inProgress: summary: Rehearsing the first round of test calls value: buildId: 425d3b75-xxxx-xxxx-xxxx-xxxxxxxxxxxx status: running round: 1 phase: testing persona: >- Ayesha Siddiqi — Lahore patient confirming tomorrow 11am, keeps asking about fees transcriptsDone: 7 transcriptsTotal: 10 failed: summary: A failed run — the reason outlives the hour of live detail value: buildId: 425d3b75-xxxx-xxxx-xxxx-xxxxxxxxxxxx assistantId: ee42ac3b-xxxx-xxxx-xxxx-xxxxxxxxxxxx status: failed phase: failed round: 0 error: 'Refine failed: LLM provider rejected the request' done: summary: Finished after three rounds, findings still open value: buildId: 425d3b75-xxxx-xxxx-xxxx-xxxxxxxxxxxx assistantId: ee42ac3b-xxxx-xxxx-xxxx-xxxxxxxxxxxx status: done phase: done round: 0 result: assistantId: ee42ac3b-xxxx-xxxx-xxxx-xxxxxxxxxxxx promptVersion: 3 passed: false chosen: happy: label: >- عائشہ صدیقی — cheerful appointment confirmation for Dr. عمران شیخ, Lahore archetype: happy path: happy turns: - speaker: agent text: >- السلام علیکم، کیا میں عائشہ صدیقی سے بات کر رہی ہوں؟ میں Shifa Clinic سے ثناء بات کر رہی ہوں۔ کل کی appointment confirm کرنی تھی، ایک منٹ بات کر سکتی ہوں؟ - speaker: caller text: >- وعلیکم السلام بھئی جی، جی میں عائشہ صدیقی ہی بول رہی ہوں۔ ہاں جی بالکل، آپ ایک منٹ بات کر لیں۔ '401': $ref: '#/components/responses/Unauthorized' '403': $ref: '#/components/responses/Forbidden' '404': description: >- No run with that id in this project — including a run that died before it could write an outcome. content: application/json: schema: $ref: '#/components/schemas/Error' example: message: No build found for 3f2a9c14-xxxx-xxxx-xxxx-xxxxxxxxxxxx error: Not Found statusCode: 404 components: schemas: BuildProgress: type: object properties: buildId: type: string format: uuid assistantId: type: string format: uuid description: >- Absent while `phase` is `authoring` through `refining`. Keep the id the start-build response gave you. status: type: string enum: - started - running - done - failed description: >- `started` lasts a moment, so the first poll usually already reads `running`. `done` and `failed` are terminal. phase: type: string enum: - authoring - generating_personas - testing - judging - refining - persisting - done - failed description: >- Step the pipeline is on: the prompt is written (`authoring`), test personas are generated, each rehearses a text-only call (`testing`), the transcripts are graded (`judging`), and anything the grader flags drives a rewrite (`refining`) before the next round. round: type: integer description: >- Which hardening round is running, `1` to `3`; `0` on frames outside the loop. persona: type: string description: >- The test persona whose rehearsal just finished. Only while `testing`. transcriptsDone: type: integer description: Rehearsals finished this round. Only while `testing`. transcriptsTotal: type: integer description: >- Rehearsals in this round. Present while `testing` and on the `judging` frame. error: type: string description: >- Why the run stopped, when `status` is `failed`. A second run against an assistant that is already building fails here rather than at the request: `A build is already in progress for this assistant`. result: $ref: '#/components/schemas/BuildOutcome' Error: type: object properties: message: description: >- One message per failed field for validation errors, a single string otherwise. oneOf: - type: string - type: array items: type: string error: type: string statusCode: type: integer BuildOutcome: type: object description: The finished build. Present only once `status` is `done`. properties: assistantId: type: string format: uuid promptVersion: type: integer description: >- Version of the prompt that was saved. It counts up once per rewrite, so anything above `1` means the grader sent it back. passed: type: boolean description: >- Whether the last round ended with nothing flagged. `false` means three rounds ran and the grader still had objections. The assistant is saved and callable either way — read the sample transcripts before putting it on a campaign. chosen: type: object description: >- The cleanest rehearsal on each path — fewest findings, longest conversation breaking ties. Either may be absent. properties: happy: $ref: '#/components/schemas/SampleTranscript' negative: $ref: '#/components/schemas/SampleTranscript' SampleTranscript: type: object properties: label: type: string description: The persona this rehearsal was run against. archetype: type: string description: The behavior the persona was built to test. enum: - happy - chatty - busy - suspicious - jailbreak - opt_out - wrong_party - mishearing - price_pusher - polite_no - maybe_later - angry path: type: string enum: - happy - negative turns: type: array items: $ref: '#/components/schemas/TranscriptTurn' TranscriptTurn: type: object properties: speaker: type: string enum: - agent - caller text: type: string responses: Unauthorized: description: A bearer key was sent but is not recognized. content: application/json: schema: $ref: '#/components/schemas/Error' example: message: invalid authorization error: Unauthorized statusCode: 401 Forbidden: description: No API key sent, or the key lacks permission for this resource. content: application/json: schema: $ref: '#/components/schemas/Error' example: message: Forbidden resource error: Forbidden statusCode: 403 securitySchemes: apiKey: type: http scheme: bearer description: Project API key (`sk_api_…`). ````