Open source audio transcription model that obscures sensitive info in realtime

Discussion: https://news.ycombinator.com/item?id=42208964