848681087e
New tool that uses webrtcvad for voice activity detection, faster-whisper for transcription, and xdotool to type into any focused window. Supports session-based listening, configurable silence threshold, and a "full stop" magic word to auto-submit. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
1.7 KiB
1.7 KiB
Project: speech-to-text tools
Speech-to-text command line utilities leveraging local models (faster-whisper, Ollama).
Environment
- Debian Bookworm, kernel 6.1, X11
- Conda env:
whisper-ollama(Python 3.10, CUDA 12.2) - mamba must be initialized before use — run:
eval "$(micromamba shell hook -s bash)" - GPU: NVIDIA (float16 capable)
- xdotool installed for keyboard simulation (X11 only)
Tools
assistant.py/talk.sh— transcribe speech, copy to clipboard, optionally send to Ollamavoice_to_terminal.py/terminal.sh— voice-controlled terminal via Ollama tool callingvoice_to_xdotool.py/dotool.sh— hands-free voice typing into any focused window (VAD + xdotool)
Testing
- To test scripts:
mamba run -n whisper-ollama python <script.py> --model-size base - Use
--model-size basefor faster iteration during development - Audio device is available — live mic testing is possible
- Test xdotool output by focusing a text editor window
Dependencies
- Conda: faster-whisper, sounddevice, numpy, pyperclip, requests, ollama
- Pip (in conda env): webrtcvad
- System: libportaudio2, xdotool
Conventions
- Shell wrappers go in .sh files using
mamba run -n whisper-ollama - All scripts set
CT2_CUDA_ALLOW_FP16=1 - Whisper model loading always has GPU (cuda/float16) -> CPU (cpu/int8) fallback
- Keep scripts self-contained (no shared module)
- Don't print output for non-actionable events
Preferences
- Prefer packages available via apt over building from source
- Check availability before recommending a dependency
- Prefer snappy/responsive defaults over cautious ones
- Avoid over-engineering — keep scripts simple and focused