Transcribe MP3

Transcribe MP3 audio files to markdown transcripts using whisper-cpp.

Requirements

• whisper-cpp installed (brew install whisper-cpp)
• ffmpeg installed (brew install ffmpeg)
• Medium model at ~/.whisper-cpp/ggml-medium.bin

Workflow

1. Identify Files

Find MP3 files to transcribe:

• If user specifies a file, use that
• Otherwise, glob for output/audio-*.mp3

Extract video_id from filename: output/audio-{video_id}.mp3