MCP server giving AI assistants terminal control, file management, process control, and code execution
Supports in-memory Python/Node.js/R execution and native Excel/PDF/DOCX handling
MIT-licensed with 5.3K+ GitHub stars — installs via npm with command blacklist for safety
Pricing: Free forever
Best for: Individuals & startups
Pros & Cons
Pros
Comprehensive local control — terminal, files, processes, and code execution in one MCP server
In-memory code execution makes quick data analysis possible without cluttering your filesystem
Large active community (5.3K+ GitHub stars) with frequent updates and responsive maintainers
Command blacklist feature provides a safety net against unintended destructive operations
Cons
Broad system access is a security risk — any MCP client with access can execute arbitrary commands
Windows support has required hotfixes — macOS and Linux remain more reliable
No built-in sandboxing or permission levels — the command blacklist is the only safety mechanism
Key Features
Terminal command execution with configurable timeouts and long-running process supportFile system operations — read, write, search, move, create directories, and get metadataSurgical diff-based file editing via edit_block — no need to rewrite entire filesIn-memory code execution for Python, Node.js, and R without saving files to diskProcess management — list running processes and kill them by PIDNative Excel, PDF, and DOCX file support for reading, writing, and text extractionFile preview UI with rendered markdown, inline images, and expandable content in Claude DesktopCommand blacklist for blocking specific operations and maintaining system safety
Desktop Commander MCP is a Model Context Protocol server that gives AI assistants full control over your local machine — terminal commands, file system operations, process management, and in-memory code execution. Built on top of MCP Filesystem Server, it extends basic file access with long-running command execution, interactive process sessions, diff-based file editing, and native support for Excel, PDF, and DOCX files.
The server provides tools across four categories. Terminal tools let you execute commands with configurable timeouts, read output from long-running processes, force-terminate sessions, and list active command sessions. Process management tools list system processes and kill them by PID. File system tools handle reading, writing, creating directories, moving files, searching by pattern, and retrieving metadata. The edit_block tool enables surgical text replacements in code files without rewriting entire contents.
Desktop Commander also includes in-memory code execution for Python, Node.js, and R — you can run scripts without saving files to disk, which is ideal for quick data analysis on CSV, JSON, or Excel datasets. A file preview UI provides visual previews in Claude Desktop with rendered markdown, inline images, and expandable content. The server has 5,300 plus GitHub stars, uses an MIT license, and installs via npm or Smithery CLI with automatic updates on restart. A command blacklist feature lets you block specific operations for safety.
It gives AI assistants full control over your local machine — terminal commands with configurable timeouts, file system operations (read, write, move, search), process management (list and kill by PID), and diff-based code editing.
Can it run code without saving files?
Yes. It includes in-memory code execution for Python, Node.js, and R. You can run scripts without saving files to disk, which is useful for quick data analysis on CSV, JSON, or Excel datasets.
Is Desktop Commander MCP safe to use?
It includes a command blacklist feature that lets you block specific operations. You configure which commands are forbidden, giving you control over what the AI agent can and cannot execute on your system.
How does it differ from the basic Filesystem MCP?
It extends Filesystem MCP with long-running command execution, interactive process sessions, diff-based editing, and native support for Excel, PDF, and DOCX files. It is a superset of the basic file server.
How do I install it?
Install via npm or Smithery CLI — it supports automatic updates on restart. It has 5,300+ GitHub stars, uses the MIT license, and works with Claude Desktop, Cursor, and other MCP clients.