How to Connect MCP Using Claude

Overview

The Model Context Protocol (MCP) is an open standard that allows Claude to securely connect to external tools, data sources, and services. This guide walks you through setting up an MCP connection using Claude.

Connecting to Brainfish specifically? See the Brainfish MCP Server article for Brainfish-specific setup instructions for Claude.ai, Cursor, and Claude Desktop.

Prerequisites

Before getting started, ensure you have:

• Claude desktop app installed (or access to Claude via API)
• An MCP-compatible server or tool you want to connect
• Basic familiarity with configuration files (JSON)

Step-by-Step Instructions

1. Open Claude Settings

Launch the Claude desktop app and navigate to Settings > Developer > MCP Servers.

2. Add a New MCP Server

Click Add Server and provide the following details:

• Name: A friendly name for your MCP connection (e.g., "My Database")
• Command: The command used to start the MCP server (e.g., npx, python, node)
• Arguments: Any arguments required by the server
• Environment Variables: Optional env vars needed by the server

3. Configure the MCP Server

MCP servers are configured via a JSON file located at:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

Example configuration:

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

4. Restart Claude

After saving your configuration, restart the Claude desktop app for the changes to take effect.

5. Verify the Connection

Once Claude restarts, you should see your MCP server listed as connected. Claude will now have access to the tools and resources exposed by that server.

Troubleshooting

Server not appearing after restart Verify your JSON configuration is valid (no syntax errors) and that the command path is correct.

Connection errors Check that any required environment variables (like API keys) are correctly set and that the MCP server is accessible from your machine.

Permission issues Ensure the MCP server binary or script has the correct execution permissions.

Further Resources

• MCP Documentation
• Available MCP Servers
• Contact your Brainfish admin if you need help configuring organization-wide MCP servers.

<!-- Reviewed by Charlie Shen, 2026-06-04 -->