koshi-vox
v1.2.6
Published
Voice-To-Text recorder with sound notifications - optimized for macOS
Downloads
22
Maintainers
Readme
██╗ ██╗ ██████╗ ███████╗██╗ ██╗██╗ ██╗ ██╗ ██████╗ ██╗ ██╗
██║ ██╔╝██╔═══██╗██╔════╝██║ ██║██║ ██║ ██║██╔═══██╗╚██╗██╔╝
█████╔╝ ██║ ██║███████╗███████║██║ ██║ ██║██║ ██║ ╚███╔╝
██╔═██╗ ██║ ██║╚════██║██╔══██║██║ ╚██╗ ██╔╝██║ ██║ ██╔██╗
██║ ██╗╚██████╔╝███████║██║ ██║██║ ╚████╔╝ ╚██████╔╝██╔╝ ██╗
╚═╝ ╚═╝ ╚═════╝ ╚══════╝╚═╝ ╚═╝╚═╝ ╚═══╝ ╚═════╝ ╚═╝ ╚═╝
░░░█ Voice-To-Text Recorder for Terminal - optimized for MacOS █░░░
░░ v1.2.6Voice-To-Text Recorder for Terminal - optimized for MacOS and M series Macs for Claude Code workflow. Record voice, get instant transcription, and paste directly into your coding sessions.
Features
- Instant Voice Recording: Press SPACE or ENTER to stop
- Fast M1-4 Transcription: Optimized for Apple Silicon
- Automatic Clipboard: Text ready to paste with Cmd+V
- Isolated Python Environment: Zero system dependencies conflicts
- Smart Uninstaller: Complete removal with installation tracking
- Customizable Alias: Choose your preferred command name
- Auto-Cleanup: Keeps only 5 most recent recordings
Installation
npx koshi-vox@latest
Dependencies
Script Installs
- faster-whisper
- soundfile
- librosa
- fastapi
- uvicorn
- nerdfonts
From Local Directory (Current Method)
# Clone/navigate to the vox directory
cd path/to/koshi-vox/
# Install globally from local directory
npm install -g .
# Set up your voice alias
setup-vox
# To uninstall run
uninstall-voxStructure
koshi-vox/
├── bin/ # Executables
│ ├── koshi-vox
│ ├── vox
│ └── vox-debug
├── scripts/ # Setup tools
│ ├── install.js
│ ├── setup-vox-alias
│ └── uninstall_koshi-vox
├── lib/ # Helpers
│ ├── install-deps
│ └── install-fonts
└── sounds/ # Audio filesRequirements
- Node.js 20+ (for npm)
- Python 3 with packages:
pip install faster-whisper soundfile - Audio Recording (macOS: sox or ffmpeg, Linux: alsa-utils)
# macOS brew install sox # Linux sudo apt install alsa-utils xclip
Usage
# Use your custom alias (set during setup)
vox
# Or use any alias you configured
my-voice-cmd
# Debug mode for troubleshooting
vox-debug
# Check recording files (auto-cleanup keeps 5 most recent)
vox-cleanup
# Remove all recording files
vox-cleanup --allRecording Workflow
- Type your voice command (
vox) - Speak your message
- Press SPACE or ENTER when done
- Text appears and is copied to clipboard
- Paste into Claude Code with Cmd+V
Configuration
Change Your Alias
# Default alias (vox)
setup-vox
# Custom alias
setup-vox myvoiceTroubleshooting
# Check if installed correctly
npm list -g @koshi-code/koshi-vox
# Reinstall if needed
npm uninstall -g @koshi-code/koshi-vox
npm install -g .
# Debug recording issues
vox-debugSystem Requirements
- macOS: Monterey+ recommended (M1-4 optimized)
- Linux: Ubuntu 20.04+ or equivalent
- Microphone: System microphone permissions required
Changelog
v1.2.6
- Fixed: VS Code terminal support - now correctly uses virtual environment Python
- Fixed: Python dependency resolution in isolated environments
- Added: Version update script for maintaining version parity across files
v1.2.5
- Initial public release
Part of the Koshi-Code ecosystem - Premium development tools for Claude Code workflows.
