microphone-linesScribe

Transcribe completed clinical recordings with medical vocabulary support and automatic keyterm resolution.

The Scribe API provides transcription for completed clinical recordings captured by the Superscribe clinical copilot. Long recordings are automatically chunked, transcribed with medical vocabulary support, and reassembled into a single transcript with recording-relative timing.

Recording Transcription

Submit a completed audio recording for transcription. The platform splits long recordings into chunks, transcribes each chunk with medical vocabulary support, and returns an assembled transcript with segment-level timing and speaker diarization.

The platform automatically resolves medical keyterms from multiple sources to improve transcription accuracy:

  • Patient context (name, primary diagnosis) when a patient_entity_id is provided

  • Service-level keyterms configured on the service

  • Workspace-level entity data

  • Request-level keyterms passed in the keyterms field

Endpoint

Method
Path
Description

POST

/v1/{workspace_id}/scribe/recordings/transcribe

Transcribe a completed clinical recording

Request

This endpoint accepts multipart/form-data with the following fields:

Field
Type
Required
Description

audio

file

Yes

Audio file to transcribe. Supported formats: WebM, OGG, WAV, MP4, MP3. Maximum size: 512 MB.

service_id

string (UUID)

Yes

Service ID for keyterm resolution

patient_entity_id

string (UUID)

No

Patient entity ID for context-aware keyterm loading

language

string

No

Language code (e.g., en, es). Defaults to patient context language or English. Maximum 16 characters.

keyterms

string

No

Comma-separated medical keyterms to boost recognition accuracy. Maximum 4000 characters.

Response

Field
Type
Description

transcript

string

Assembled clinical transcript text

confidence

number or null

Duration-weighted transcription confidence score

duration_seconds

number

Total recording duration in seconds

model

string

Transcription engine identifier

language

string

Language used for transcription

chunk_count

integer

Number of audio chunks processed

keyterm_count

integer

Number of normalized medical keyterms used for vocabulary boosting

chunks

array

Per-chunk transcription details

segments

array

Utterance-level segments with recording-relative timing

Chunk object

Field
Type
Description

index

integer

Zero-based chunk index

start_seconds

number

Chunk start time relative to recording start

end_seconds

number

Chunk end time relative to recording start

text

string

Transcript text for this chunk

confidence

number or null

Transcription confidence for this chunk

duration_seconds

number or null

Decoded audio duration for this chunk

Segment object

Field
Type
Description

text

string

Segment transcript text

start_seconds

number or null

Segment start time relative to recording start

end_seconds

number or null

Segment end time relative to recording start

confidence

number or null

Transcription confidence for this segment

speaker

integer or null

Speaker index (when diarization is available)

Error Responses

Status
Description

404

Service not found in the workspace

413

Audio file exceeds the 512 MB size limit

415

Unsupported audio content type

422

Empty audio file

504

Transcription timed out

Permissions

Requires service.view permission on the workspace.

Physician Review

The Scribe API also supports physician review workflows for encounter documentation generated by the clinical copilot. These endpoints write human review events for the encounter entity.

Method
Path
What it does

POST

/v1/{workspace_id}/scribe/encounters/{encounter_id}/icd10/approve

Approve a suggested ICD-10 code

POST

/v1/{workspace_id}/scribe/encounters/{encounter_id}/icd10/reject

Reject a suggested ICD-10 code

POST

/v1/{workspace_id}/scribe/encounters/{encounter_id}/soap/edit

Replace a SOAP section with physician-edited content

POST

/v1/{workspace_id}/scribe/encounters/{encounter_id}/finalize

Finalize the encounter for downstream sync

Approve ICD-10

Response:

Reject ICD-10

Response:

Edit SOAP

section must be one of subjective, objective, assessment, or plan.

Response:

Finalize

Finalize does not require a request body.

For real-time microphone streaming, transcript events, SOAP updates, ICD-10 suggestions, and session lifecycle events, see Clinical Copilot.

Last updated

Was this helpful?