Building AI-Powered Tools with Claude + MCP

What You Will Build

In this tutorial, you will connect Claude Desktop to a real API using FuzeMCP. By the end, Claude will be able to query your API directly through natural language.

Prerequisites

• A FuzeMCP account
• A REST API endpoint (or use our demo API)
• Claude Desktop installed

Step 1: Create a FuzeMCP Project

1. Log in to your FuzeMCP dashboard
2. Click New Project
3. Name your project and enter your API base URL
4. Configure authentication if your API requires it

Step 2: Define Your Tools

Each tool maps an API endpoint to an MCP tool definition. For example:

{
  "name": "get_product",
  "description": "Fetch product details by reference",
  "inputSchema": {
    "type": "object",
    "properties": {
      "reference": {
        "type": "string",
        "description": "Product reference code"
      }
    },
    "required": ["reference"]
  }
}

Step 3: Get Your Endpoint

After defining your tools, FuzeMCP generates a hosted MCP endpoint:

https://api.fuzemcp.dev/mcp/your-project-id

Step 4: Connect Claude Desktop

Add the endpoint to Claude Desktop's configuration:

{
  "mcpServers": {
    "my-api": {
      "url": "https://api.fuzemcp.dev/mcp/your-project-id",
      "headers": {
        "Authorization": "Bearer your-mcp-api-key"
      }
    }
  }
}

Restart Claude Desktop. You should now see your tools available.

Step 5: Ask Claude to Use Your API

Try asking Claude:

"Find product with reference BO8711827031109"

Claude will call your MCP tool and return the result.

Next Steps

• Add more tools to your project
• Explore FuzeMCP documentation for advanced features
• Connect Cursor, VS Code, or other MCP clients