Model Context Protocol servers

What is Model Context Protocol servers?

This repository is a collection of reference implementations for the Model Context Protocol (MCP), along with references to community-built servers and additional resources. It showcases how MCP can give Large Language Models (LLMs) secure, controlled access to tools and data sources. Each server is implemented with either the TypeScript MCP SDK or the Python MCP SDK.

How to use Model Context Protocol servers?

Clone the repository and refer to each server’s directory (e.g., src/) for individual setup and configuration instructions. The servers are designed to be run independently, typically requiring Node.js or Python, and appropriate API keys or local service connections.

Key features of Model Context Protocol servers

• Secure file operations with configurable access controls
• Web and local search using Brave’s Search API
• Read-only database access with schema inspection (PostgreSQL, Sqlite)
• Browser automation and web scraping via Puppeteer
• Knowledge graph–based persistent memory system
• Time and timezone conversion capabilities

Use cases of Model Context Protocol servers

• Allow AI assistants to query and interact with GitHub repositories
• Enable LLMs to fetch and convert web content for efficient processing
• Give models controlled access to local files and directories
• Integrate AI agents with Google Drive file search and retrieval

FAQ from Model Context Protocol servers

What is the Model Context Protocol?

MCP is an open protocol that standardizes how applications provide context and tools to LLMs, enabling secure, controlled interactions.

Which SDKs are used to build these servers?

The reference servers are built with the TypeScript MCP SDK or the Python MCP SDK.

How do I add a new MCP server to this collection?

The repository encourages community contributions; you can submit a pull request to add your server to the “Third-Party Servers” or “Reference Servers” sections. Follow the existing structure and alphabetical order.

Are there any runtime dependencies?

Each server may depend on Node.js or Python, plus specific SDKs and external APIs (e.g., Brave Search, PostgreSQL client). See the individual server’s README for details.

Where does data reside when using these servers?

Data storage depends on the server: the Memory server maintains a local knowledge graph, Filesystem accesses the local file system, and cloud-based servers (e.g., GitHub, Google Drive) interact with remote APIs.