Drop a call recording in, get a speaker-labelled transcript out. Runs against the local WhisperX service.
Nothing transcribed yet.