Model Context Protocol (MCP)

An open standard that lets AI assistants connect to apps, files and data through "MCP servers".

by Anthropic (open standard) No ratings yet

What it is

The Model Context Protocol (MCP) is an open standard for connecting AI assistants to external tools and data. Think of it as a universal plug: an MCP server exposes abilities (search my files, create a ticket, query a database), and any MCP client (Claude, ChatGPT, Cursor, VS Code and others) can use them.

Key ideas

  • Servers offer tools (actions), resources (data to read) and prompts (templates).
  • Clients are the AI apps that connect to servers and let the model call those tools.
  • Local servers run on your computer; remote servers run online and usually sign in with OAuth.
  • Thousands of servers exist for GitHub, Slack, Google Drive, databases, browsers and more.

Getting started

  • In Claude apps, add ready-made connectors from Settings → Connectors.
  • In Claude Code, add a server from the terminal:
Terminal
claude mcp add my-server -- npx -y some-mcp-server
  • In Claude Desktop, local servers can also be configured in the claude_desktop_config.json file.

Full walkthrough: Connect an MCP server to Claude.

Tips

  • Start with official servers from the service you want to connect.
  • Give each server the minimum permissions it needs.

Warning

An MCP server can read data and take actions for the AI. Install only trusted servers, and be careful when a tool reads content from the web or email, as that content may contain instructions meant to trick the AI.

Guides for Model Context Protocol (MCP)

Intermediate 15 min

Connect an MCP server to Claude

Give Claude new abilities, such as reading your files or using GitHub, by connecting it to MCP servers in the app or in Claude Code.

Uses: Model Context Protocol (MCP), Claude, Claude Code

Comments 0

Share tips, prompts that worked for you, or questions.

No comments yet. Be the first to share a tip.