How to Deploy autoMate with Docker and Connect Any AI Client via MCP
Step-by-step: run the autoMate container, connect OpenClaw, Claude Desktop, Cursor or Cline over MCP, and drive your real Chrome with the browser extension.
In this tutorial you'll deploy autoMate — the self-hosted computer-use agent — with Docker, then connect it to any AI client (OpenClaw, Claude Desktop, Cursor, or Cline) over MCP in a few minutes.
🚀 Want to deploy autoMate yourself?
Docker configs, system requirements, and installation guides — all on one page.
View autoMate Tool Page →Step 1 — Run the container
The official image ships with Python 3.12, Playwright Chromium, and all browser dependencies preinstalled. Desktop tools are skipped inside the container (no display), but browser automation and every data tool work out of the box:
docker run -d --name automate \ -p 8765:8765 \ -v automate-data:/data \ ghcr.io/yuruotong1/automate:latest
Open http://localhost:8765 — the setup wizard walks you through picking a model, pasting an API key, and wiring up an AI client. All state (notes, files, memory) lives in the /data volume in SQLite, encrypted with Fernet.
Step 2 — Connect your AI client via MCP
After install, open Settings → Connect to AI clients and click "Copy install text". You get a single blob with the URL + token pre-filled, plus per-client sections. Three ways to use it:
| Method | What you do |
|---|---|
| Manual | Edit your client's config file by hand using the generated text |
| Paste to another AI | "Cursor, here's autoMate, set it up for me" — the text is written so an AI can configure itself |
| OpenClaw bundle-mcp | Paste the section into your OpenClaw config |
Once the client picks up the new server it gains all of autoMate's tools — search.find, notes.read, files.list, audio.transcribe — plus a top-level automate tool that runs autoMate's own agent loop on demand.
Step 3 — Drive your real browser
For full computer-use (clicking through your actual Chrome with your logged-in sessions), install the browser extension and point it at ws://localhost:8765/api/extension/ws — the default. Your real browser becomes a tool your agent can use, while your data stays local.
🚀 Want to deploy autoMate yourself?
Docker configs, system requirements, and installation guides — all on one page.
View autoMate Tool Page →That's it — a private computer-use agent with persistent memory, running on your own hardware. Deploy it today and give any AI client real hands.