Introduction

MCP (Model Context Protocol) servers extend Claude Code's capabilities by connecting it to external tools and data sources. Think of MCP servers as plugins — each one gives Claude new abilities, like querying a database, searching documentation, or interacting with a third-party API.

This folder contains comprehensive documentation and examples for MCP server configurations and usage with Claude Code.

Without MCP servers, Claude Code can only read files, run shell commands, and search your codebase. With them, you can give Claude direct access to services like GitHub, Slack, Postgres, or any custom API — no manual copy-pasting required. This matters because it turns Claude from a code assistant into a fully integrated development partner.

MCP servers are configured in your settings.json file. Here's what a basic configuration looks like:

{
  "mcpServers": {
    "my-server": {
      "command": "npx",
      "args": ["-y", "@example/mcp-server"],
      "env": {
        "API_KEY": "your-key-here"
      }
    }
  }
}

This tells Claude Code to launch the MCP server using npx, pass it any required arguments, and set environment variables like API keys. Once configured, the server's tools become available to Claude automatically — you can reference them by name in your prompts.

You can configure MCP servers at two levels:

• Project-level — stored in .claude/settings.json within your repo, shared with your team
• User-level — stored in ~/.claude/settings.json, available across all your projects

Key takeaway: MCP servers are the primary way to give Claude Code new capabilities beyond its built-in tools. Start with one server that solves a real problem in your workflow, then add more as needed.