Core REST API

`GET /meetings/{id}/audio-url`

Returns a short-lived signed URL for direct audio playback from Cloud Storage.


Auth	`bearerAuth`
Cache-Control	`private, no-store`
Response	`200 { url: string, expires_at: string }`
Errors	`404` meeting not found; `404` meeting has no audio

`POST /meetings`

Creates a meeting. Set source_type: "recording" for live recordings.


Auth	`bearerAuth`
Idempotency	`Idempotency-Key: <uuid>` header required
Echo suppression	`X-Client-Id: <uuid>` header
Response	`201 Created` with full `Meeting` body

`PATCH /meetings/{id}`

Partially updates meeting metadata. Used by ML to set the headline after post-meeting processing.


Auth	`bearerAuth` or `serviceAuth`
Echo suppression	`X-Client-Id: <uuid>` header
Response	`200` with full `Meeting` body
Side effects	Broadcasts `EntityChanged { entity: meeting, action: updated }`

`POST /meetings/{id}/speaker-labels`

Action endpoint that associates an AssemblyAI speaker label with a known Person. Triggers a bulk person_id update across all segments sharing that label and enriches the Person's voice embedding.


Auth	`bearerAuth`
Idempotency	`Idempotency-Key: <uuid>` header required
Echo suppression	`X-Client-Id: <uuid>` header
Response	`200 { updated_count: integer }`
Errors	`404` meeting not found; `422` `person_id` does not exist; `422` `speaker_label` not found in this meeting's transcript

Request body:

{
  "speaker_label": "Speaker A",
  "person_id": "uuid"
}

Side effects:

All transcript_segments with speaker_label = "Speaker A" receive person_id
Person's voice_vector updated from aggregated segment feature_vector values
EntityChanged { entity: transcript_segment, action: updated } broadcast (echo-suppressed)
EntityChanged { entity: person, action: updated } broadcast (echo-suppressed)

`PUT /meetings/{id}/synthesis`

Atomically replaces the meeting's AI-generated synthesis. Called by ML after post-meeting processing.


Auth	`serviceAuth`
Echo suppression	`X-Client-Id: <uuid>` header
Response	`204 No Content`
Side effects	Broadcasts `EntityChanged { entity: meeting, action: updated }`

`POST /meetings/{id}/talking-points`

Creates a talking point for a meeting. Called by ML to deliver live insights per finalised segment.


Auth	`serviceAuth`
Idempotency	`Idempotency-Key: <uuid>` header required
Echo suppression	`X-Client-Id: <uuid>` header
Response	`201 Created` with full `TalkingPoint` body
Side effects	Broadcasts `TalkingPointEvent` on WebSocket

`POST /transcriptions/{transcriptionId}/segments`

Appends transcript segments to a transcription. Used on the live path.


Auth	`serviceAuth`
Response	`204 No Content`
Side effects	Broadcasts `TranscriptSegmentEvent` on WebSocket for each final segment

`PUT /transcriptions/{transcriptionId}/segments`

Atomically replaces all segments for a transcription. Used by the post-meeting batch worker.


Auth	`serviceAuth`
Response	`204 No Content`
Errors	`404` transcription not found; `409` transcription is still in `processing` state
Side effects	Deletes existing segments, inserts new set, broadcasts `EntityChanged { entity: transcript_segment }`

`DELETE /meetings/{meetingId}/tasks?source=system`

Deletes all system-generated tasks for a meeting. Used by the post-meeting worker when re-extracting tasks.


Auth	`serviceAuth`
Response	`204 No Content`

Core REST API

On this page