Transcribe Feishu voice/audio messages into text using local faster-whisper. Use when the user sends a Feishu voice message, asks to convert Feishu audio to text, or wants to turn local audio into Chinese text for further analysis.
Use this skill to convert Feishu voice or audio files into text with local faster-whisper.
/home/node/clawd/vendor/faster-whisper./lzcapp/document/fzj/openclaw/offline-wheels/faster-whisper-smallzh) and CPU int8 inference.Transcribe a local audio file:
{baseDir}/scripts/transcribe-audio-faster-whisper.py /path/to/audio.wav
Specify model alias and language explicitly:
{baseDir}/scripts/transcribe-audio-faster-whisper.py /path/to/audio.wav small zh
For Feishu voice/audio messages, first download the message resource to local disk, then pass the saved file path into the script.
The script returns JSON with:
text: merged transcription textsegments: timestamped segmentsdetected_languageaudio_pathmodel_sourcescripts/transcribe-audio-faster-whisper.pyMain local transcription entrypoint.