📌Overview

Purpose: mcp-get is designed to facilitate the installation and management of Model Context Protocol (MCP) servers, enabling seamless interactions between LLM applications and a variety of external data sources and tools.

Overview: The Model Context Protocol (MCP) provides a standardized interface for integrating language models with diverse applications, allowing developers to connect LLMs with necessary context from various data sources. The mcp-get tool simplifies the process of deploying such MCP servers, making it easier for developers to enhance their applications.

Key Features:

Package Management: Easily install, update, and uninstall MCP servers tailored for different tools, such as GitHub, Slack, and databases, facilitating effective integration between LLMs and data sources.
Cross-Platform Compatibility: While primarily developed for Linux and macOS environments, there is potential support for Windows, promoting wider accessibility for various development platforms.

mcp-get

A command-line tool for installing and managing Model Context Protocol (MCP) servers.

About Model Context Protocol

The Model Context Protocol (MCP) is an open protocol that enables seamless integration between LLM applications and external data sources and tools. Whether you're building an AI-powered IDE, enhancing a chat interface, or creating custom AI workflows, MCP provides a standardized way to connect LLMs with the context they need.

Learn more about MCP at modelcontextprotocol.io

What Packages Can You Install?

This tool helps you install and manage MCP servers that connect Claude to various data sources and tools, including:

Development Tools: GitHub, GitLab
Communication Tools: Slack
Search & Data: Brave Search, Google Maps
Database Systems: PostgreSQL, SQLite
Web Automation: Puppeteer
Cloud Storage: Google Drive

Prerequisites

Node.js (version 14 or higher) for Node.js-based MCP servers
Python (version 3.10 or higher) for Python-based MCP servers
Claude Desktop app (for local MCP server usage)

Note: This tool has not been thoroughly tested on Windows systems yet. While it may work, you might encounter some issues. Contributions to improve Windows compatibility are welcome!

Usage Examples

Install a Package

npx @michaellatman/mcp-get@latest install @modelcontextprotocol/server-brave-search

Sample output:

Installing @modelcontextprotocol/server-brave-search...
Installation complete.

List Packages

npx @michaellatman/mcp-get@latest list

Sample output:

📦 Available Packages
Found 11 packages

@modelcontextprotocol/server-brave-search │ MCP server for Brave Search API integration │ Anthropic, PBC (https://anthropic.com) │ MIT
@modelcontextprotocol/server-everything   │ MCP server that exercises all the features of the MCP protocol │ Anthropic, PBC (https://anthropic.com) │ MIT
...

Uninstall a Package

npx @michaellatman/mcp-get@latest uninstall @modelcontextprotocol/server-brave-search

Sample output:

Uninstalling @modelcontextprotocol/server-brave-search...
Uninstallation complete.

Update the Tool

The tool automatically checks for updates when running commands. You can also manually update:

npx @michaellatman/mcp-get@latest update

Sample output:

Updating mcp-get...
Update complete.

Contributing

We welcome contributions to the project! If you would like to contribute, please:

Fork the repository and create a new branch for your feature or bugfix.
Write tests for your changes and ensure all existing tests pass.
Submit a pull request with a clear description of your changes.

License

This project is licensed under the MIT License.

Contact Information

If you have any questions or need help, feel free to reach out via GitHub Issues: https://github.com/michaellatman/mcp-get/issues

Adding Your Own MCP Server to the Registry

There are two ways to add your MCP server to the registry:

Option 1: Manual Package Addition

If you want to maintain your own package:

Create Your MCP Server:
- Develop your MCP server according to the MCP protocol specifications at https://modelcontextprotocol.io
- Publish it as either an NPM package (installable via npm) or a Python package (installable via uvx)
Add Your Package to the Package Folder: Add your server as a JSON file in the packages/ directory:
- For regular packages, use packages/your-package-name.json
- For scoped packages, use packages/scope--package-name.json (double hyphens between scope and name)
Example JSON structure:
```
{
  "name": "your-package-name",
  "description": "A brief description of your MCP server",
  "vendor": "Your Name or Organization",
  "sourceUrl": "URL to the source code repository",
  "homepage": "URL to the homepage or documentation",
  "license": "License type (e.g., MIT)",
  "runtime": "node | python",
  "environmentVariables": {
    "SOME_API_KEY": {
      "description": "Description of what this key is for",
      "required": true
    }
  }
}
```
Important notes:
- The name field must be the exact resolvable package name in npm or pip.
- The runtime field specifies how your package should be installed:
  - Use "runtime": "node" for npm packages.
  - Use "runtime": "python" for Python packages installable via uvx.
- Always include an environmentVariables object (can be empty {} if none required).
Validate Your Package: Run the PR check to validate your package:
```
npm run pr-check
```
Submit a Pull Request: Fork this repository, add your package file, and submit a PR.

Option 2: Community Servers Repository

If you prefer not to manage package deployment and distribution:

Fork Community Repository:
- Fork https://github.com/mcp-get/community-servers
- This repository follows the same structure as the official MCP servers.
Add Your Server:
- Add your implementation to the src directory.
- Follow existing patterns and structure.
- Include necessary documentation and tests.
Submit a Pull Request:
- Submit your PR to the community servers repository.
- Once merged, your server will be automatically added to the registry.

Both options require following the MCP protocol specifications at https://modelcontextprotocol.io. Choose the option that best suits your needs:

Option 1 if you want full control over your package distribution.
Option 2 if you want to avoid managing package deployment and distribution.