Anibon Local Audio Transcription

Overview & Triggers

Use when YouTube has no subtitles or auto-captions for the target video.

1. Audio Extraction

Download audio stream as mono 16kHz WAV file (required for whisper.cpp miniaudio and to prevent >4GB WAV header overflow on long streams):

yt-dlp -x --audio-format wav --exec "ffmpeg -i {} -ar 16000 -ac 1 -c:a pcm_s16le audio_16k.wav && rm {}" "VIDEO_URL" -o "temp_audio.%(ext)s"

Or convert existing audio:

ffmpeg -i audio.wav -ar 16000 -ac 1 -c:a pcm_s16le audio_16k.wav

2. Local Transcription