Have you explored using speaker diarization and speaker identification, given that pyannote etc. takes this approach?
I'm curious given your decision to capture speaker names from the screen. I see the merits during desktop recording, but I can also see how this limits utility when trying to offer the same functionality across desktop and other scenarios (e.g. in-person meetings, audio uploads etc.)
We already support diarization in the Desktop Recording SDK by capturing the meeting platform’s speaker-change events, so you get a diarized transcript plus precise “speaker started talking” timestamps out of the box. We also support voice-signature diarization via third-party STT providers for participants calling in from the same room
For in-person meetings and audio uploads, this is on our roadmap and in development. More to come on this!