Desktop Commander
Lets Claude run terminal commands, search and edit files on your computer through MCP.wonderwhy-er/DesktopCommanderMCP
- Stars
- 10.2k
- Forks
- 1.3k
- Licence
- MIT
- Language
- JavaScript
- Last update
- 4 days ago
- Type
- Dev tool
What it is
Desktop Commander is an MCP server that connects AI assistants such as Claude Desktop to your computer. It lets the assistant run terminal commands, manage processes, search folders, and edit files, including targeted text replacements. It is built on the MCP Filesystem Server and adds search and replace editing on top.
- Search a code folder for text and make targeted replacements across several files
- Run a development server and check its output in pages without flooding the chat
- Analyze a CSV, JSON, or Excel file by asking the assistant questions about it
- Create a PDF from markdown or extract text from an existing PDF
- Run short Python, Node.js, or R snippets in memory without saving files
Great for
- Developers who want an AI assistant to run commands and edit files in a local project
- People using Claude Desktop who want file and terminal access without paying per API token
- Users who work with CSV, Excel, PDF, or Word files and want an assistant to read or edit them
- Anyone running long processes such as dev servers or database sessions that need monitoring
Not ideal for
- People who need a strict security boundary, since the README says it is not a sandbox and only adds guardrails like a command blocklist
- Non-technical users who are uneasy granting an AI tool access to their filesystem and shell
- Teams that need a managed, audited enterprise deployment, as the README describes local-only history and manual setup for some options
How to get it
- 1
Install with a package manager
You'll need: Node.js
Install and configure for Claude Desktop
npx @wonderwhy-er/desktop-commander@latest setup - 2
Run it
- Run the setup command in a terminal
- Restart Claude Desktop if it is running
- Ask Claude to perform a file or terminal task in a folder you choose
- 3
Get the most out of it
- Use the file preview UI and markdown editor for rendered previews in Claude Desktop
- Use process output pagination with offset and length controls to avoid context overflow
- Use the negative offset option to read the end of log files, like tail
- Use the Remote MCP option to connect from ChatGPT or Claude web
Downloads
Strengths
- Installs with a single npx command and updates automatically on restart for most install options
- Covers files, terminal, processes, Excel, PDF, and DOCX in one server
- Keeps a local record of tool calls on the machine running it
Trade-offs
- The README states it is not a sandbox, so granting it shell access carries real risk
- Some features and the Docker option depend on setup steps and external tools the README only partly describes
- Option 5 (local checkout) requires manual git updates
How it compares
| Project | Stars | Type | Level | Updated | Best for |
|---|---|---|---|---|---|
| DesktopCommanderMCP (this) | 10.2k | Dev tool | beginner | 4 days ago | Developers who want an AI assistant to run commands and edit files in a local project |
| codebase-memory-mcp | 48k | Dev tool | intermediate | 4 days ago | Developers who use an MCP-capable AI coding agent on a large codebase |
| claude-context | 12.7k | Dev tool | beginner | 4 months ago | Developers working in large codebases where loading everything into an agent's context is impractical |
| context7 | 64.1k | AI model / agent | beginner | 4 days ago | Developers who use AI coding agents such as Claude, Cursor, or VS Code for libraries that change often |
| page-agent | 29.6k | Library | intermediate | 6 days ago | Developers adding an AI copilot to their own web app |
| ECC | 295k | Dev tool | intermediate | 4 days ago | Developers who already use Claude Code and want a shared workflow for planning, testing and review |
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.