/transcripts endpoint

Overview

Speech to Text

Facial Physiological Engine

Vocal Acoustic Engine

Neurobehavioral Construct Computer

/api/transcripts is the REST batch transcription surface — upload a recorded encounter, get a structured transcript + instrument scores back. Same engine, same DSM-5-TR vocabulary bias, same C-SSRS safety check as /sessions and /transcribe.

Upload

POST /api/sessions/{id}/recordings
Content-Type: multipart/form-data

Limits: 60 minutes audio, 150 MB file. See Audio Configuration.

Trigger transcript

POST /api/sessions/{id}/transcripts
Content-Type: application/json

{
  "language": "en-US",
  "diarization": true,
  "speakers": 2,
  "instruments": ["phq-9", "gad-7", "c-ssrs"],
  "vocabulary": { "formulary": [...], "programs": [...] }
}

Poll status

GET /api/sessions/{id}/transcripts/{transcript_id}/status

Status: processing → completed | failed.

Retrieve

GET /api/sessions/{id}/transcripts/{transcript_id}

Returns the same structured shape /sessions returns at finalize — transcript, diarization, instrument scores, codes, SOAP draft.

Async batch accuracy may be slightly lower than the real-time endpoints because batch loses interactive context (no in-the-moment vocabulary biasing). Pass the practice vocabulary on every request for parity.

What’s next

/sessions — live capture

Stateful diarized session.

/transcribe — dictation

Between-visit dictation.

Audio configuration

Formats and limits.

Ambient async quickstart

End-to-end async walkthrough.

​Upload

​Trigger transcript

​Poll status

​Retrieve

​What’s next

/sessions — live capture

/transcribe — dictation

Audio configuration

Ambient async quickstart

Upload

Trigger transcript

Poll status

Retrieve

What’s next