Transcript resources ^BETA

One of the key resources produced by a Meeting Agent is the transcript.

Example Response #

{
  "meeting_agent": {
    "sub": "agt_69286a86d432e0005a6eb9ca"
  },
  "transcript": {
    "summary": "A summary of the meeting as a whole",
    "languages": [
      {
        "language": "en"
      }
    ],
    "paragraphs": [
      {
        "speaker": {
          "identifier": "A",
          "name": "Garry Shutler"
        },
        "start": 2.72,
        "end": 7.6,
        "sentences": [
          {
            "text": "This is a simple example of the transcription format.",
            "start": 2.72,
            "end": 7.6
          }
        ]
      }
    ]
  }
}

Response Parameters #

meeting_agent.sub required  #

The sub of the Meeting Agent that generated the transcript.

transcript.summary required  #

A String providing a plain text summary of the meeting.

transcript.languages[] required  #

An array of languages detected for the meeting.

By default we transcribe by detecting the single dominant language of the meeting, so a single value is returned. Where the Meeting Agent was requested with a language of detect_multi — see provisioning and scheduling — a value is returned for each language detected across the meeting.

Integrators should tolerate zero values present, as we may be unable to detect transcribable content.

transcript.languages[].language required  #

A String containing an IETF BCP 47 language tag for a language detected in the meeting, such as en or fr.

transcript.paragraphs[].speaker.identifier required  #

A String consistently representing the speaker throughout the transcript.

transcript.paragraphs[].speaker.name optional  #

A String consistently representing the name of the speaker throughout the transcript.

It is highly likely this parameter will be provided, but it cannot be guaranteed and so an integration should tolerate its absence.

transcript.paragraphs[].start required  #

A Float representing number of seconds since the start of the transcript the paragraph began being spoken.

transcript.paragraphs[].end required  #

A Float representing number of seconds since the start of the transcript the paragraph stopped being spoken.

transcript.paragraphs[].sentences[].start required  #

A Float representing number of seconds since the start of the transcript the sentence began being spoken.

transcript.paragraphs[].sentences[].end required  #

A Float representing number of seconds since the start of the transcript the sentence stopped being spoken.

transcript.paragraphs[].sentences[].text required  #

A String detailing what was said for the sentence.