V16 - General AI Agent with browser automation, secure credential vault, workflow management, email checking, and full desktop control. Let AI control your computer.
npm install v16.ai> AI that actually does things
V16 connects your local machine to a powerful AI agent that can control your entire computer - mouse, keyboard, screen capture, browser automation, code execution, and more.
``bash`
npm install -g v16.ai
This will also install Python dependencies for desktop control.
1. Sign up at v16.ai
2. Get your token from v16.ai/settings
3. Connect your machine:
`bash`
v16 connect --token YOUR_TOKEN
4. Start using V16 - Go to the dashboard and tell the agent what to do
š Quick Start Guide | Full Documentation
bash
Check Gmail unread count
curl -X POST http://localhost:3000/api/local-agent/email/unread \
-H "x-user-id: demo" -d '{"provider": "gmail"}'Search the web
curl -X POST http://localhost:3000/api/local-agent/search \
-H "x-user-id: demo" -d '{"query": "latest AI news"}'Execute a saved workflow
curl -X POST http://localhost:3000/api/local-agent/workflows/{id}/execute \
-H "x-user-id: demo"
`$3
- "Check my Gmail and summarize unread emails"
- "Search for 'best React frameworks 2026' and extract top 5 results"
- "Fill out this contact form with my saved credentials"
- "Open Chrome, navigate to Hacker News, and screenshot the top stories"
- "Take a screenshot and describe what you see"
- "Open VS Code and create a new React component"Commands
`bash
v16 connect # Connect to V16
v16 connect --token XXX # Connect with specific token
v16 test # Test all capabilities (CLI, browser, desktop)
v16 login # Get token instructions
v16 --help # Show all commands
`Desktop Control Capabilities
| Command | Description |
|---------|-------------|
|
screenshot | Capture full screen |
| mouse-move | Move cursor to x,y coordinates |
| mouse-click | Click at position |
| mouse-double-click | Double click |
| mouse-right-click | Right click |
| mouse-drag | Drag from point A to B |
| mouse-scroll | Scroll up/down |
| keyboard-type | Type text |
| keyboard-press | Press single key (Enter, Escape, etc.) |
| keyboard-hotkey | Key combo (Cmd+C, Ctrl+V, etc.) |
| clipboard-read | Read clipboard content |
| clipboard-write | Write to clipboard |
| window-list | List all open windows |
| window-focus | Focus/activate an app |Requirements
- Node.js 18+
- Python 3.8+ with pip
- macOS, Windows, or Linux
$3
- pyautogui - Mouse and keyboard control
- mss - Fast screen capture
- pyperclip - Clipboard access
- Pillow - Image processing$3
On macOS, you need to grant accessibility permissions:
1. Go to System Preferences ā Security & Privacy ā Privacy
2. Select "Accessibility" from the left sidebar
3. Add Terminal (or your terminal app) to the listEnvironment Variables
| Variable | Description |
|----------|-------------|
|
V16_TOKEN | Your authentication token |
| V16_SERVER | Custom server URL (default: https://api.v16.ai) |Security
- All communication is encrypted (HTTPS/WSS)
- Only authenticated commands from your V16 account are executed
- You control what the agent can access
- Disconnect anytime with
v16 logout`Note: The agent has full control over your computer. Only connect machines you trust.
- Website: v16.ai
- Discord: discord.gg/AnBesyRQ4F
- Issues: github.com/v16ai/v16/issues
MIT