Get Session Transcript

Retrieve conversation transcript from completed Ambient session

GET /api/v1/ambient/session/{ambient_session_id}/transcript

cURL

cURL

curl --request GET \
  --url https://sdp.suki.ai/api/v1/ambient/session/<ambient_session_id>/transcript \
  --header 'sdp_suki_token: <sdp_suki_token>' \
  --header 'sdp_provider_id: <sdp_provider_id>'

200

{
  "final_transcript": [
    {
      "end_offset": {
        "hours": 0,
        "minutes": 6,
        "nanos": 80000000,
        "seconds": 42
      },
      "end_time": "2024-12-04T09:40:48.792948332Z",
      "lang_id": "en",
      "recording_id": "c9d59aa8-cd48-4f5a-be81-5d0c9d2a5885",
      "start_offset": {
        "hours": 0,
        "minutes": 6,
        "nanos": 80000000,
        "seconds": 42
      },
      "start_time": "2024-12-04T09:40:42.393948332Z",
      "transcript": "The patient has shown an allergy to pollen",
      "transcript_id": "01JE8GP4RTHH0KDEGRSTRVPMGH"
    }
  ]
}

Updated:

The response will now include the new lang_id field within the payload. The lang_id field indicates the language in which the transcript was sent.

Use this endpoint to get the full transcript for a specified ambient session after it has completed.

For a full list of language codes and their corresponding languages, refer to the Language code reference section.

• Call this endpoint after the session status is completed.

• For guidance on displaying the detected language and building multilingual support into your product UI, refer to Multilingual support.

• For the complete retrieval flow, refer to Get the generated clinical note.

Code examples

Python

import requests

ambient_session_id = "123dfg-456dfg-789dfg-012dfg"
url = f"https://sdp.suki-stage.com/api/v1/ambient/session/{ambient_session_id}/transcript"

headers = {
    "sdp_suki_token": "<sdp_suki_token>",
    "sdp_provider_id": "<sdp_provider_id>"
}

response = requests.get(url, headers=headers)

if response.status_code == 200:
    transcript_data = response.json()
    print("Transcript:")
    for transcript in transcript_data.get("final_transcript", []):
        print(f"Transcript ID: {transcript.get('transcript_id')}")
        print(f"Recording ID: {transcript.get('recording_id')}")
        print(f"Language: {transcript.get('lang_id')}")
        print(f"Transcript: {transcript.get('transcript')}")
        print(f"Start Time: {transcript.get('start_time')}")
        print(f"End Time: {transcript.get('end_time')}")

        # Start offset (relative to beginning of audio)
        start_offset = transcript.get('start_offset', {})
        if start_offset:
            print(f"Start Offset: {start_offset.get('hours')}h {start_offset.get('minutes')}m {start_offset.get('seconds')}s")

        # End offset (relative to beginning of audio)
        end_offset = transcript.get('end_offset', {})
        if end_offset:
            print(f"End Offset: {end_offset.get('hours')}h {end_offset.get('minutes')}m {end_offset.get('seconds')}s")

        print("---")
else:
    print(f"Failed to get transcript: {response.status_code}")
    print(response.json())

TypeScript

const ambientSessionId = '123dfg-456dfg-789dfg-012dfg';
const response = await fetch(
  `https://sdp.suki-stage.com/api/v1/ambient/session/${ambientSessionId}/transcript`,
  {
    headers: {
      'sdp_suki_token': '<sdp_suki_token>',
    'sdp_provider_id': '<sdp_provider_id>'
    }
  }
);

if (response.ok) {
  const transcriptData = await response.json();
  console.log('Transcript:');
  transcriptData.final_transcript?.forEach((transcript: any) => {
    console.log(`Transcript ID: ${transcript.transcript_id}`);
    console.log(`Recording ID: ${transcript.recording_id}`);
    console.log(`Language: ${transcript.lang_id}`);
    console.log(`Transcript: ${transcript.transcript}`);
    console.log(`Start Time: ${transcript.start_time}`);
    console.log(`End Time: ${transcript.end_time}`);

    // Start offset (relative to beginning of audio)
    if (transcript.start_offset) {
      const { hours, minutes, seconds } = transcript.start_offset;
      console.log(`Start Offset: ${hours}h ${minutes}m ${seconds}s`);
    }

    // End offset (relative to beginning of audio)
    if (transcript.end_offset) {
      const { hours, minutes, seconds } = transcript.end_offset;
      console.log(`End Offset: ${hours}h ${minutes}m ${seconds}s`);
    }

    console.log('---');
  });
} else {
  const error = await response.json();
  console.error(`Failed to get transcript: ${response.status}`, error);
}

Authorizations

sdp_suki_token · string · header · required

Suki access token (suki_token) from Login or Register. Expires after one hour.

Headers

sdp_provider_id · string

Optional for standard partners.

Required for:

• Bearer authentication. Use the same provider_id returned by the Login or Register API.
• Single Auth Token authentication. Include the same provider_id on every request as sdp_provider_id.

Example:

"provider-123"

Path Parameters

ambient_session_id · string · required

UUID for the ambient session. Use the ambient_session_id returned from Create Ambient Session, or the UUID you supplied in that request.

Response

200

application/json

Request succeeded.

Response body for the /session/{ambient_session_id}/transcript endpoint

final_transcript · object[]

Collection of transcripts for the ambient session

Show child attributes