Transcribes MP3, WAV, and M4A audio files into text using hardware-accelerated MLX Whisper on Apple Silicon.
This skill enables Claude to perform high-speed, local audio transcription directly on Apple Silicon hardware. By leveraging the MLX Whisper framework and the whisper-large-v3-turbo model, it allows users to convert audio files from formats like MP3, WAV, and M4A into text without relying on external cloud APIs. It is particularly useful for developers and researchers who need private, fast, and cost-effective speech-to-text processing for meetings, interviews, or media content directly through the Claude Code interface.
主な機能
01Local hardware acceleration via Apple Silicon GPU and Neural Engine
02Per-file timing statistics for performance monitoring
03Support for common audio formats including MP3, WAV, and M4A
04Batch processing capabilities for entire directories of audio files
050 GitHub stars
06High-accuracy transcription using the whisper-large-v3-turbo model
ユースケース
01Batch processing large folders of voice notes for content indexing
02Converting meeting recordings or interviews into searchable text documents
03Privacy-sensitive transcription of audio data without cloud uploads
