KoljaB / RealtimeSTT

A robust, efficient, low-latency speech-to-text library with advanced voice activity detection, wake word activation and instant transcription.
MIT License
2.09k stars 190 forks source link

fix: clear text storage and audio frames on audio queue clear #144

Closed SlickTorpedo closed 2 weeks ago

SlickTorpedo commented 2 weeks ago

This is NOT tested very much. Look it over and make sure it's not going to induce any downstream breaking changes!

SlickTorpedo commented 2 weeks ago

This clears it for the on_realtime_transcription_stabilized call because otherwise it's cached and won't clear properly.