Local speaker diarization using MLX Whisper (macOS) or faster-whisper (Linux/CUDA) and Pyannote
pip install voxscriber