The ability to receive live transcripts from a webhook, including speaker diarization metadata would be super useful.