transcribe_media
ChatGPTAudio and video transcription using the Evidano transcription model, covering 100+ languages, descriptive speaker labels, cross-talk, and two or more languages in one recording. Pass the temporary download URL of the user's uploaded audio/video file. Optional prompt: at most 2000 characters, for vocabulary, speaker names, language and formatting guidance. Default prompt: Transcribe this without timestamps with descriptive speaker labels in native language. Returns the transcript as a long string, with actual source start and end timestamps in separate fields; inline timestamps are omitted by default. For video, Evidano processes the video itself, including its audio and visual speaker context. Processing limits and audio-activity detection are enforced by the server. If the result is trimmed, preserve the continuation link exactly.






