WhisperLiveKit – Real-time & local speech-to-text, translation, and speaker diarization. With server & web UI.
SMRTR summary
WhisperLiveKit enables real-time speech transcription locally with speaker identification, offering advanced features beyond basic Whisper models. The system uses state-of-the-art research like SimulStreaming and Streaming Sortformer to intelligently process speech chunks while maintaining context, and supports translation to over 100 languages.
SMRTR provides this summary for quick context. The original article belongs to Github.
Read the original article