Transcribe
Transcribes audio files to verbatim text, optionally with speaker diarization, by running a bundled CLI and validating the output. Use when the user asks to transcribe a recording, meeting, interview, or podcast, or to label who said what.