What can I control with this MCP server?

You can capture screenshots, move and click the mouse at specific coordinates, and send keyboard input to automate interactions with desktop applications and GUIs.

How do I install and run computer-use-mcp?

Install via npm with `npx computer-use-mcp@1.8.0`. The server uses stdio transport, so it communicates with your MCP client through standard input/output streams.

What operating systems does this support?

The registry metadata does not specify OS support. Check the GitHub repository (io.github.domdomegg/computer-use-mcp) for platform compatibility details.

Can I use this to automate web browsers?

Yes, you can take screenshots of browser windows and automate mouse and keyboard input to interact with web applications, though dedicated browser automation tools may be more efficient for that use case.

Is there a way to record and replay sequences of actions?

The registry data does not mention recording or replay features. Refer to the project documentation on GitHub to see if such capabilities are available or planned.

MCP ServerSTDIOOfficialv1.8.0

Computer Use MCP Server

Automate computer tasks by controlling screenshots, mouse, and keyboard input through an MCP server. Ideal for developers building AI agents that need to interact with desktop applications and GUIs.

io.github.domdomegg/computer-use-mcp

Try MCP with AI models View source

Hosted URL

Local install

Transport

STDIO

Auth

No auth required

What the Computer Use MCP server does

How models use it and what it is built for.

Automate computer tasks by controlling screenshots, mouse, and keyboard input through an MCP server. Ideal for developers building AI agents that need to interact with desktop applications and GUIs.

Connect to Computer Use

Local install — runs as a subprocess.

npx computer-use-mcp@1.8.0

Resources

Where to find authoritative docs and source for Computer Use.

Example prompts for Computer Use

Paste any of these into Agent Studio after connecting Computer Use.

Take a screenshot and describe what's currently on screen
Click the button at coordinates 450, 300 then type my password
Move the mouse to the top-left corner and take another screenshot
Automate filling out a form by taking screenshots between each keyboard input

Computer Use MCP server — FAQ

Common questions about connecting and running Computer Use.

What can I control with this MCP server?
You can capture screenshots, move and click the mouse at specific coordinates, and send keyboard input to automate interactions with desktop applications and GUIs.
How do I install and run computer-use-mcp?
Install via npm with `npx computer-use-mcp@1.8.0`. The server uses stdio transport, so it communicates with your MCP client through standard input/output streams.
What operating systems does this support?
The registry metadata does not specify OS support. Check the GitHub repository (io.github.domdomegg/computer-use-mcp) for platform compatibility details.
Can I use this to automate web browsers?
Yes, you can take screenshots of browser windows and automate mouse and keyboard input to interact with web applications, though dedicated browser automation tools may be more efficient for that use case.
Is there a way to record and replay sequences of actions?
The registry data does not mention recording or replay features. Refer to the project documentation on GitHub to see if such capabilities are available or planned.

Skip the local setup — run MCP in your browser

MCP Playground runs 10,000+ hosted MCP servers — GitHub, Linear, Notion, Stripe, Sentry and more — across Claude, GPT, Gemini, DeepSeek and 40+ AI models. Compare model answers side-by-side, save agent presets, share runs. Zero install.

Open Agent Studio