servers
Model Context Protocol Servers
claude mcp add modelcontextprotocol-servers
How to use
This repository hosts a collection of reference MCP servers that demonstrate Model Context Protocol features and SDK usage. It includes reference implementations such as Everything, Fetch, Filesystem, Git, Memory, Sequential Thinking, and Time, each illustrating tools and data access patterns that enable LLMs to interact with prompts, resources, and external capabilities in a controlled way. Use these references to study MCP concepts, SDK integration, and common patterns for building secure, tool-enabled agents. To experiment, review the individual server directories, understand the available tools, and reference the generated MCP metadata and schemas used by the MCP registry.
Each reference server exposes a defined set of capabilities (e.g., content retrieval, file operations, repository access, or reasoning workflows). You can adapt the concepts to your own environment by inspecting how prompts, resources, and tools are wired together, how authorization and resource access are managed, and how results are returned to the LLM in a structured, auditable fashion.
How to install
Prerequisites:
- Git
- Node.js and npm (for TypeScript/JavaScript examples) or Python (for Python SDK examples), depending on the specific reference server you want to run.
- Access to the MCP SDKs described in the repository (as applicable).
Step-by-step:
-
Clone the repository git clone https://github.com/modelcontextprotocol/servers cd servers
-
Choose a reference server to run (e.g., Everything, Fetch, Filesystem, Git, Memory, Sequential Thinking, Time) and install its dependencies following its language-specific guidelines. For example, if the server is a TypeScript/Node.js implementation:
- Navigate to the server directory, e.g.: cd src/everything
- Install dependencies: npm install
- Build (if required): npm run build
-
Run the server with its typical command, and configure environment variables as needed. Common patterns include:
- Node.js: npm start or node dist/server.js with appropriate env vars
- Python: python -m <module> or uvicorn <app>:app for FastAPI-based examples
-
Register or verify the MCP metadata as needed with your MCP registry or local testing harness, ensuring the server exposes the expected endpoints and tool interfaces.
Additional notes
Notes and tips:
- These are reference implementations intended for educational use and experimentation. They are not guaranteed to be production-ready.
- Review environment variables and security considerations in each server directory to understand how access to tools and data is controlled.
- If you plan to run multiple servers locally, consider isolating them with distinct ports and careful credential management.
- The MCP registry and the archived servers provide additional context and examples beyond the in-repo references.
Related MCP Servers
iterm
A Model Context Protocol server that executes commands in the current iTerm session - useful for REPL and CLI assistance
mcp
Octopus Deploy Official MCP Server
furi
CLI & API for MCP management
editor
MCP Server for Phaser Editor
DoorDash
MCP server from JordanDalton/DoorDash-MCP-Server
mcp
MCP сервер для автоматического создания и развертывания приложений в Timeweb Cloud