jarvis-launcher 1.2.0

Two claps → JARVIS speaks → launches your workspace

JARVIS Launcher

Clap twice → JARVIS speaks → your workspace opens automatically.

Detects two consecutive claps via microphone and launches Claude Code, Cursor, and a YouTube song of your choice. Runs as a background app with a system tray icon on Linux, macOS, and Windows.

Install

pip install jarvis-launcher
jarvis-launcher

No system dependencies required on any platform.

Download binary (no Python needed)

Grab the latest binary from Releases.

Platform	File
Linux	jarvis-launcher-linux
macOS	jarvis-launcher-macos
Windows	jarvis-launcher-windows.exe

Usage

Start the app

jarvis-launcher

A green tray icon appears. The terminal shows mic levels in real-time:

[████████░░░░░░░░░] 🔔🔔

• Green icon = listening
• Yellow icon = sound detected
• Red icon = triggered
• Gray icon = paused

Right-click the tray icon to pause/resume, open settings, or quit.

CLI commands

jarvis config          # Show current configuration
jarvis set sensitivity 0.2  # Change a setting
jarvis voices          # List available TTS voices
jarvis open-settings   # Open config in editor
jarvis test-mic        # Test microphone levels

Configuration

Settings live in ~/.jarvis-launcher/config.json (created on first run). Open it via the tray menu or edit directly:

{
  "sensitivity": 0.15,
  "required_claps": 2,
  "open_claude_code": true,
  "open_cursor": true,
  "open_youtube": true,
  "youtube_url": "https://www.youtube.com/watch?v=v2AC41dglnM",
  "voice": "en-GB-RyanNeural"
}

Sensitivity tuning:

Value	Environment
0.08	Very quiet room
0.15	Normal office (default)
0.25	Noisy environment

Changes take effect on the next clap trigger — no restart needed.

How it works

• Streams microphone audio in 1024-sample chunks at 44100 Hz
• Each chunk is checked for clap signature: high RMS + energy concentrated in the 1–8 kHz band (>25% of total)
• Adaptive threshold adjusts to background noise automatically
• Two claps within 3 seconds trigger the launch sequence

License

MIT