Transcript Model#
While most users will end up browsing and searching transcripts using the CDP frontend tooling, we also need a standardized method for storing the transcript data for both pipeline processing and enabling a stable API for research and archival purposes.
Below you will find an example transcript.
For more specific details please see the
object definition documentation.
Example Transcript#
{
"generator": "EvaGen -- Lib Version: 0.0.0",
"confidence": 0.93325,
"session_datetime": "2021-01-10T15:00:00",
"created_datetime": "2024-03-01T17:53:57.199132",
"sentences": [
{
"index": 0,
"confidence": 0.9,
"start_time": 0.0,
"end_time": 1.0,
"words": [
{
"index": 0,
"start_time": 0.0,
"end_time": 0.5,
"text": "hello",
"annotations": null
},
{
"index": 1,
"start_time": 0.5,
"end_time": 1.0,
"text": "everyone",
"annotations": null
}
],
"text": "Hello everyone.",
"speaker_index": 0,
"speaker_name": "Eva Maxfield Brown",
"annotations": null
},
{
"index": 1,
"confidence": 0.95,
"start_time": 1.0,
"end_time": 2.0,
"words": [
{
"index": 0,
"start_time": 1.0,
"end_time": 1.5,
"text": "hi",
"annotations": null
},
{
"index": 1,
"start_time": 1.5,
"end_time": 2.0,
"text": "all",
"annotations": null
}
],
"text": "Hi all.",
"speaker_index": 1,
"speaker_name": "Isaac Na",
"annotations": null
}
],
"annotations": {
"sections": [
{
"name": "Call to Order",
"start_sentence_index": 0,
"stop_sentence_index": 2,
"generator": "Eva Maxfield Brown",
"description": null
}
]
}
}