Connect MCP Servers to Give Claude Code Access to Your Tools and Services

Out of the box, Claude Code reads files and runs shell commands. MCP (Model Context Protocol) servers let you plug in anything else — databases, APIs, documentation, project management tools — so Claude can use them natively.

{
  "mcpServers": {
    "my-docs": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-filesystem", "./docs"]
    }
  }
}

Add that to .claude/settings.json (project-level) or ~/.claude/settings.json (global), and Claude gains new tools it can call during your session.

The MCP ecosystem has servers for common needs:

{
  "mcpServers": {
    "github": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-github"]
    },
    "postgres": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-postgres"]
    },
    "memory": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-memory"]
    }
  }
}

Once connected, Claude sees the MCP tools alongside its built-in tools. Ask "check the database for users created today" and it queries directly. Ask "create a GitHub issue for this bug" and it does — no copy-pasting between windows.

You can also build custom MCP servers for your team's internal tools. Any service with an API can become an MCP server that Claude calls naturally during a conversation.

A few practical setups:

• Database server — Claude queries your schema and data directly
• Documentation server — Claude searches your internal docs for answers
• Monitoring server — Claude checks dashboards and alerts while debugging

MCP servers turn Claude Code from a code editor into a command center for your entire development stack.

via Claude Code