SuperAPI AI Directory
Partner with us中文AI newsFavorites
All toolsAI developer platformsmcpm.sh

mcpm.sh

MCPM (Model Context Protocol Manager) is an open-source CLI tool for managing MCP (Model Context Protocol) servers. It simplifies management through a global configuration model: a server only needs to be installed once, can be organized by tags through virtual configurations (Profiles), and integrated into any MCP client. Core features include discovering servers from a central registry, direct execution, public sharing, server updates, client integration management (such as Claude Desktop, Cursor, Windsurf), and non-interactive automation support for AI agents. MCPM is built with Python and Click, supports multiple installation methods such as brew, pipx, and uv, and provides a rich command-line interface.

Open website

About this tool

MCPM is an open-source CLI tool for managing Model Context Protocol (MCP) servers, maintained by pathintegral-institute. It aims to simplify the global configuration of MCP servers, allowing users to install a server once and then organize it through virtual configurations (Profiles) and integrate it into any MCP client.

Main purpose: manage the installation, discovery, configuration, execution, and updating of MCP servers, with support for multi-client integration.

Main capabilities:

  • Global server management: install once, use anywhere.
  • Virtual configurations (Profiles): organize servers by tags to adapt to different workflows.
  • Server discovery: browse and install servers from the MCP registry (mcpm.sh/registry).
  • Direct execution: run servers via stdio or HTTP for easy testing or integration.
  • Public sharing: share servers through a secure tunnel.
  • Server updates: check for and apply updates based on git, NPX, and binary servers.
  • Client integration: manage the configuration of clients such as Claude Desktop, Cursor, and Windsurf.
  • AI agent friendly: provide a non-interactive CLI and environment variable support, as well as an llm.txt guide, for easy automation.
  • Beautiful CLI: rich text formatting and an interactive interface.
  • Usage analytics: monitor server usage and performance.

How to use: MCPM v2.0 uses a global configuration model. Installation methods include brew, pipx, uv, and others. Basic command examples: install a server, list installed servers, serve over HTTP, create a configuration and add a server, run a configuration. The CLI provides interactive and non-interactive modes, and full automation can be enabled through environment variables.

Applicable scenarios: suitable for developers, AI agent developers, and users who need to manage multiple MCP servers, to simplify server configuration, integration, and update processes and improve work efficiency.

Limitations and notes:

  • The supported MCP clients currently explicitly include Claude Desktop; other clients (such as Cursor and Windsurf) are mentioned in the documentation but may be in development or partially supported.
  • Version v2.0 has been fully released, and some features (such as server update management, GitHub release binary updates, advanced monitoring, Docker execution, etc.) are listed as future enhancements and are not yet available.
  • DemoThe video shows outdated v1 commands, and a new demo is still being prepared.
  • For specific pricing, costs, and platform compatibility details, please refer to the official documentation.

Related tools

Browse category