Why does the API return the original transcript after web edits, and how do I get SRT subtitles?

The public API uses the original transcript. Saving text or speaker edits on the website does not overwrite it. Export from the website for the edited version. For subtitles, use the transcription download endpoint on a completed task with format=srt; available formats are listed in capabilities under downloadFormats. SRT includes only timed segments. If there is only untimed text, the subtitle file may be empty, so check the task details as well.

GET /sound_clone/api/v1/stt/transcriptions/{transcriptionId}/download?format=srt