whisper.cpp

Local speech-to-text with CUDA acceleration. Runs OpenAI Whisper models locally for real-time transcription.