Dispatch Mode

Let your agent control your computer — take screenshots, click buttons, type text, and navigate any app on your Mac or PC.

How It Works

Your agent runs on a dedicated server 24/7. Dispatch Mode creates a secure tunnel between your agent and your personal computer. When you run the relay, your agent can see your screen and interact with any application — just like a remote assistant sitting at your desk.

Quick Start

Easiest way: Go to instaclaw.io/settings and click "Connect Your Computer". It downloads a small file — double-click it and you're connected.

Or from your terminal:

npx @instaclaw/dispatch

You'll be asked for a pairing code (shown on your Settings page). Enter it and the relay connects automatically via encrypted WebSocket.

Requirements

  • Node.js 18 or later
  • Pro or Power tier subscription
  • macOS: Grant Accessibility and Screen Recording permissions to your terminal app (Terminal, iTerm, Warp, etc.)
  • Windows: No special permissions needed
  • Linux: X11 display server (Wayland not yet supported)

What Your Agent Can Do

  • Take screenshots of your screen
  • Click at any position
  • Type text via keyboard
  • Press key combos (Cmd+C, Ctrl+V, etc.)
  • Drag and drop between positions
  • Scroll in any direction
  • List open windows
  • Open and control any desktop application
  • Batch actions — execute multiple actions in one round-trip (click, type, press Enter) for 2-3x faster task completion

Modes

Supervised (default)

Every action the agent wants to take is shown in your terminal. Press Enter to approve or 'n' to deny. Screenshots are auto-approved.

Autonomous

Auto-approves most actions. Dangerous actions (passwords, delete, purchases) still require your confirmation. Enable with:

npx @instaclaw/dispatch --autonomous

Speed

Dispatch v0.5 introduced action batching and optimized screenshots for significantly faster computer control.

  • Batch actions: Your agent plans multiple steps (click, type, press Enter) and executes them in a single round-trip instead of one at a time — 2-3x faster for multi-step tasks
  • WebP screenshots: 50-60% smaller than the previous JPEG format, reducing transfer time without losing visual quality
  • Smart verification: Your agent only takes screenshots at decision points, not after every keystroke — cutting vision costs by up to 70%

A typical 20-step task that previously took 1-3 minutes now completes in 25-45 seconds.

macOS Permissions

macOS requires two permissions for computer control:

  1. Accessibility — allows mouse clicks and keyboard input.
    System Settings → Privacy & Security → Accessibility → enable your terminal app
  2. Screen Recording — allows screenshots.
    System Settings → Privacy & Security → Screen Recording → enable your terminal app

Note: macOS Sequoia may re-prompt for Screen Recording permission monthly. This is an Apple security feature, not an InstaClaw issue.

Security

  • All communication is encrypted via TLS (WebSocket Secure)
  • Authentication uses your unique gateway token — only your agent can send commands
  • Supervised mode lets you approve every action before it executes
  • Dangerous actions (passwords, deletions, purchases) always require confirmation, even in autonomous mode
  • Press Ctrl+C at any time to immediately disconnect and stop all agent control

FAQ

Does my agent have access to my computer when the relay isn't running?

No. Your agent can only control your computer while the relay is actively running in your terminal. Close it and the connection is severed immediately.

What data does the relay send to my agent?

Screenshots (as compressed WebP images) and command results (success/failure). No files, passwords, or browsing history are ever sent.

Can I use this on multiple computers?

The relay connects one computer at a time. Run it on whichever machine you want your agent to control.

Why does my agent need to control my computer?

For tasks that require your installed apps, logged-in sessions, or local files — things your agent can't access from its server. Examples: editing a Figma file, filling out a form in your browser, organizing files on your desktop.