Computer MCP
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
Computer MCP
Computer MCP, or pc-mcp, is a small server that gives AI agents controlled access to your computer through the Model Context Protocol. With it running, an agent can browse and edit files, search, organize folders, create and extract zip files, run terminal commands, start and control background processes and move files between your PC and the sandbox it works in.
When it starts, the server creates an access token and opens a public address through a Cloudflare tunnel, downloading cloudflared the first time if it is not installed. It then shows the address and the token to paste into any MCP client and opens a status page in the browser. If you prefer, the tunnel can be turned off to use only the local network, or pointed to your own address from ngrok or a named Cloudflare tunnel.
Since this hands a lot of power to an agent, there are a few ways to limit it. By default it can see the whole computer, but you can restrict access to specific folders, start in read only mode, which blocks writing, moving, deleting, the terminal and imports, or keep the files open and only turn off the terminal. Download and upload links expire after two hours by default.
How to use
You need Node.js 18.17 or newer. On Windows, double click Iniciar pc-mcp.bat, or the read only version next to it. On Mac or Linux run iniciar.sh, which also builds the project the first time. After that, copy the address and the token into your MCP client.
If you prefer to do it by hand, run npm install, npm run build and npm start. Running it with the --help option lists everything you can change, like the port, the tunnel mode, the allowed folders and the default shell. Options can be saved for the next runs and can also come from environment variables such as PC_MCP_TOKEN and PC_MCP_PORT.
Related MCP server: Desktop Commander MCP
Development time
Computer MCP was built in one hour.
License
GNU General Public License version 3. The full text is in the LICENSE file.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Use your Mac, Windows or Linux computer from ChatGPT, Claude or Codex: files, commands, documents.
Connect AI assistants to your approved computer for files, commands, browser and desktop tasks.
Work on your own computers from Claude or ChatGPT: run commands, edit and search files.
Give your AI hands on every computer you own. Remote control of your Windows, Mac and Linux machines from one ChatGPT, Claude or Cursor chat — even from your phone. Ask it to find out why the office server is crawling, update every PC in the house, sort out your mom's laptop from another city, or ship the site from your work machine while you're on a train. It runs commands, reads files and takes a screenshot when it needs to see the screen. You decide what each computer allows, and you can cut any of them off in a second.
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceEnables AI assistants to search, update, and manage files, run terminal commands, and automate workflows on a user's computer.1,287,096 npmMIT
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to search, update, manage files, and run terminal commands on a computer, with remote access via Streamable HTTP and OAuth 2.0 for integration with ChatGPT and web agents.1,287,096 npm1MIT
- FlicenseAqualityDmaintenanceEnables an LLM agent to search files, list directories, read and write UTF-8 text files, move or delete paths, list terminal processes, and launch a terminal in a chosen folder.8-
- AlicenseNot gradedqualityDmaintenanceEnables AI clients to search, read, edit, and manage files and directories, execute and control terminal commands and processes, and perform code or text searches and replacements. It also supports in-memory code execution and interactive process sessions for tasks like data analysis and long-running command management.1,287,096 npm2MIT