MCP Reference Servers

MCP Reference Servers is a collection of locally run examples for developers building connections between AI applications and external tools or data. Maintained by the MCP steering group, the servers demonstrate the protocol and its SDKs. They're educational implementations, so developers should assess security requirements before using them in production.

The collection covers file access, repository work and other tasks an AI client can request:

• Filesystem provides file operations with configurable access controls. Git supports reading, searching and manipulating repositories.
• Memory stores persistent information in a knowledge graph. Sequential Thinking supports problem-solving through sequences of thoughts.
• Fetch retrieves web content and converts it for LLM use. Time handles time and timezone conversions.
• Everything is a test server that demonstrates MCP prompts, resources and tools.

The servers work through an MCP client, with Claude Desktop given as one example. MCP also has support in ChatGPT, Visual Studio Code and Cursor. The repository includes TypeScript and Python implementations, giving developers concrete examples to adapt for their own integrations.

Local file and repository servers expose data on the machine where they run; Fetch accesses external websites. Running a server locally doesn't determine where the connected AI client processes that data. The collection focuses on reference implementations, while the MCP Registry lists published servers.

Code is transitioning from MIT to Apache 2.0: new contributions use Apache 2.0, while older contributions without relicensing consent retain MIT terms. Documentation is separately CC-BY-4.0.