MCP Setup Guide

Standard MCP Configuration

All MCP-compatible AI assistants use the same configuration format. Add this to your AI assistant's MCP settings:

{
  "mcpServers": {
    "brick-directory": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-fetch", "https://api.brick.directory/connect"]
    }
  }
}

After adding this configuration, restart your AI assistant to activate the connection.

Need platform-specific help? Check the tabs above for detailed instructions for your specific AI assistant.

Claude Web (claude.ai) or Desktop - Custom Connector

For Claude Web UI or Desktop, use the custom connector feature:

---

Claude Code - MCP Configuration

For Claude Code, run the following command:

claude mcp add --transport http brick-directory https://api.brick.directory/connect

Then run /mcp command from within Claude Code. Select brick-directory from the list and select Authenticate. You will be prompted for the login in the browser window and once authenticated, you will be able to use Brick Directory from your Claude Code.

ChatGPT Configuration

Note: ChatGPT's MCP support is currently in beta and may require specific setup steps.

For ChatGPT Desktop with MCP support, use the generic MCP configuration above.

Alternative: If you're using ChatGPT's Deep Research feature, you may need to configure remote MCP servers through their specific interface.

Check the latest OpenAI Community for updated MCP configuration instructions.

Gemini CLI Configuration

Add this to your Gemini CLI settings.json file using SSE transport:

{
  "mcpServers": {
    "brick-directory": {
      "url": "https://api.brick.directory/connect",
      "timeout": "10m"
    }
  }
}

Or use the command line:

gemini mcp add brick-directory --url https://api.brick.directory/connect

Then verify the configuration with:

gemini mcp list

See the official Gemini MCP documentation for more details.