Model Context Protocol Servers
by modelcontextprotocol
This repository provides reference implementations for the Model Context Protocol (MCP), showcasing how LLMs can securely access tools and data sources. Each server is implemented with either the Typescript or Python MCP SDK.
Last updated: N/A
What is Model Context Protocol Servers?
This repository is a collection of reference server implementations for the Model Context Protocol (MCP). MCP enables Large Language Models (LLMs) to securely and controllably access external tools and data sources.
How to use Model Context Protocol Servers?
The servers can be used as-is or as templates for building custom integrations. Typescript-based servers can be run directly with npx. Python-based servers can be run using uvx or pip. To use a server, configure it within an MCP client like Claude Desktop, specifying the command and arguments to start the server.
Key features of Model Context Protocol Servers
Reference implementations of MCP servers
Demonstrates versatility and extensibility of MCP
Secure and controlled access to tools and data sources for LLMs
Implemented with Typescript and Python MCP SDKs
Includes a variety of servers for different use cases
Use cases of Model Context Protocol Servers
Providing LLMs with access to file systems
Enabling LLMs to interact with databases
Allowing LLMs to perform web searches
Giving LLMs access to Git repositories
Integrating LLMs with third-party services like GitHub, Google Drive, and Slack
FAQ from Model Context Protocol Servers
What is the Model Context Protocol (MCP)?
What is the Model Context Protocol (MCP)?
MCP is a protocol that allows Large Language Models (LLMs) to securely and controllably access external tools and data sources.
What are reference servers?
What are reference servers?
Reference servers are example implementations of MCP servers that demonstrate how to use the protocol and the SDKs.
How do I use the servers in this repository?
How do I use the servers in this repository?
The servers can be used as-is or as templates for building custom integrations. Typescript-based servers can be run directly with npx. Python-based servers can be run using uvx or pip. To use a server, configure it within an MCP client like Claude Desktop, specifying the command and arguments to start the server.
What is the difference between reference servers and third-party servers?
What is the difference between reference servers and third-party servers?
Reference servers are maintained by the Model Context Protocol team and are intended to demonstrate MCP features and the SDKs. Third-party servers are maintained by companies or individuals outside of the Model Context Protocol team.
Are community servers safe to use?
Are community servers safe to use?
Community servers are untested and should be used at your own risk. They are not affiliated with or endorsed by Anthropic.