Uses faster whisper, returns word probabilities

alesaccoia / VoiceStreamAI

Near-Realtime audio transcription using self-hosted Whisper and WebSocket in Python/JS

MIT License

736 stars 106 forks source link

Closed alesaccoia closed 10 months ago

alesaccoia commented 10 months ago

Updated demo with word probabilities, uses faster whisper as default ASR, refactored tests