Transcription
Turn a recording into text with timestamps.
Tools1
media__transcribe9 credits / minuteTranscribe speech in an audio or video file to text, with speaker labels and optional word timings. `media` is a Drive path or file id, or a public https URL: WAV, MP3, AAC/M4A, FLAC, Ogg, WebM or MOV, up to 60 minutes and 100 MB. Send the whole conversation in one call — speaker labels are consistent only within one answer. The language is detected; `language` is only a hint. The transcript is the text answer. Speakers are marked inline (`<|speaker:0|>`), with emotion and vocal-event cues such as `[laughter]` unless `tag_audio_events` is false. `timestamps: true` adds word-level `segments` and timed `speaker_turns` to structuredContent. Long recordings take minutes to process — prefer no client-side timeout for this call. Billed per second of the whole recording (silence included). Failures never charge.Transcribe speech in an audio or video file to text, with speaker labels and optional word timings. `media` is a Drive path or file id, or a public https URL: WAV, MP3, AAC/M4A, FLAC, Ogg, WebM or MOV, up to 60 minutes and 100 MB. Send the whole conversation in one call — speaker labels are consistent only within one answer. The language is detected; `language` is only a hint. The transcript is the text answer. Speakers are marked inline (`<|speaker:0|>`), with emotion and vocal-event cues such as `[laughter]` unless `tag_audio_events` is false. `timestamps: true` adds word-level `segments` and timed `speaker_turns` to structuredContent. Long recordings take minutes to process — prefer no client-side timeout for this call. Billed per second of the whole recording (silence included). Failures never charge.
