$npx -y skills add SafeAI-Lab-X/ClawKeeper --skill openai-whisperLocal speech-to-text with the Whisper CLI (no API key).
| 1 | # Whisper (CLI) |
| 2 | |
| 3 | Use `whisper` to transcribe audio locally. |
| 4 | |
| 5 | Quick start |
| 6 | |
| 7 | - `whisper /path/audio.mp3 --model medium --output_format txt --output_dir .` |
| 8 | - `whisper /path/audio.m4a --task translate --output_format srt` |
| 9 | |
| 10 | Notes |
| 11 | |
| 12 | - Models download to `~/.cache/whisper` on first run. |
| 13 | - `--model` defaults to `turbo` on this install. |
| 14 | - Use smaller models for speed, larger for accuracy. |