Voice
Every Voice tool, with the scope it needs, its arguments and the API operation it calls.
3 tools
Get voice transcription jobs
Read onlyatlas_voice_transcription_jobs_get
Fetch the current status/result of a bounded server-side ASR job. Completed results include speaker-turn diarization handoff and summary-ready transcript metadata.
- Scope
ai:read- Calls
- Example prompt
- Check transcription job asr_123.
Arguments
| Field | Type | Required | Description |
|---|---|---|---|
jobId | string | Yes | Voice ASR job id returned by atlas_voice_transcription_jobs_start. |
Start voice transcription jobs
Changes dataatlas_voice_transcription_jobs_start
Start a bounded server-side ASR job for audio bytes. Returns a job receipt with status, expiry, requested mime/locale, and later result/error fields; Atlas does not persist raw audio in the job record.
- Scope
ai:write- Calls
- Example prompt
- Start a transcription job for this WebM audio.
Arguments
| Field | Type | Required | Description |
|---|---|---|---|
audioBase64 | string | Yes | Base64-encoded audio bytes (up to 8 MB decoded). Accepts standard or URL-safe encoding. |
mime | string | Yes | Source audio mime type, e.g. audio/webm, audio/mp4, audio/wav. |
locale | string | No | Optional BCP-47 locale hint, e.g. en-US, hi-IN. |
Create voice transcriptions
Changes dataatlas_voice_transcriptions_create
Transcribe audio bytes via the configured server-side ASR provider. Returns { text, segments, confidence, provider }. When ASR is not configured the response carries provider="not_configured" and an empty transcript - no audio leaves Atlas.
- Scope
ai:write- Calls
- Example prompt
- Transcribe this voice memo and pull out the action items.
Arguments
| Field | Type | Required | Description |
|---|---|---|---|
audioBase64 | string | Yes | Base64-encoded audio bytes (up to 8 MB decoded). Accepts standard or URL-safe encoding. |
mime | string | Yes | Source audio mime type, e.g. audio/webm, audio/mp4, audio/wav. |
locale | string | No | Optional BCP-47 locale hint, e.g. en-US, hi-IN. |