The canonical reference collection of MCP (Model Context Protocol) servers maintained by Anthropic. Includes production-ready implementations for filesystem access, GitHub integration, Slack, PostgreSQL, SQLite, Google Maps, Brave Search, Puppeteer browser automation, memory tools, and 20+ more. Every MCP client — Claude Desktop, Cursor, Windsurf, Cline — can connect to these servers.
MCP Official Servers is the reference implementation of the Model Context Protocol — Anthropic’s open standard for connecting AI agents to external tools and data sources. Think of it as “USB-C for AI tools” — any MCP-compatible client can connect to any MCP server.
This collection includes 20+ production-ready servers covering the most common integration needs: file system access, Git/GitHub, databases (PostgreSQL, SQLite), web search (Brave), browser automation (Puppeteer), communication (Slack), mapping (Google Maps), and persistent memory.
{
"mcpServers": {
"filesystem": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-filesystem", "/path/to/allowed/dir"]
},
"github": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-github"],
"env": { "GITHUB_PERSONAL_ACCESS_TOKEN": "<your-token>" }
}
}
}
User: "Find all open issues labeled 'bug' in my repo and summarize them"
Agent → GitHub MCP Server:
- list_issues(owner: "user", repo: "myapp", labels: ["bug"], state: "open")
- Returns 12 issues with titles, descriptions, and labels
Agent: "You have 12 open bugs. The top priorities are:
1. Memory leak in WebSocket handler (#142) — reported 3 days ago
2. Auth token expiry not handled (#138) — affects 5 users"
AI agents that work well with MCP Official Servers.
The fast, Pythonic way to build MCP servers — like FastAPI but for the Model Context Protocol.
Secure cloud sandboxes for AI agents to execute code — isolated environments with built-in safety.
Automatically expose FastAPI endpoints as MCP tools — turn any API into an AI-callable tool in one line.