H2h2o
Documents

Returns speaker-segment timestamps for an audio/video transcript document.

Returns exact per-segment start/end times (seconds), speaker, and text for an audio or video transcript document, captured from speaker diarization at ingest. Returns an empty list for non-transcript documents and documents ingested before this data was captured.

get/documents/{document_id}/audio_timestamps

Request

Path parameters

document_idstring required

Id of the audio/video document.

Response

Successful operation

startnumber double required

Segment start time in seconds.

endnumber double nullable

Segment end time in seconds. Null if unknown (rare edge cases).

speakerstring required

Assigned/display speaker name for the segment: the name inferred by the LLM (e.g. "Alice") when available, otherwise the raw diarization label. Empty if not diarized.

original_speakerstring nullable

Raw diarization speaker label for the segment (e.g. "SPEAKER_01", "SPEAKER_-1" for the unknown speaker). Preserved alongside speaker because the LLM may map two distinct diarization speakers onto the same name; this exposes what the diarization model actually distinguished. Empty for non-diarized/scene segments.

textstring required

Transcribed text of the segment.

typestring nullable

Segment kind for video transcripts (e.g. scene vs speech); absent for audio.

Changes

No recorded changes to this operation across all 1 revision of this API.