MCP Hatchery 🥚

A CLI tool to scaffold Model Context Protocol (MCP) servers with support for:

• JavaScript: Using @modelcontextprotocol/sdk with Netlify/Vercel hosting
• Python: Using FastMCP framework with optional FastMCP Cloud hosting

Installation

npm install -g mcp-hatchery

Usage

mcp-hatchery create my-server

Or use directly with npx:

npx mcp-hatchery create my-mcp-server

You'll be prompted for a project name and description. The tool generates a complete MCP server with example tools ready for customization.

Project Types

JavaScript (Bare Bones)

MCP server using the official JavaScript SDK:

• Local stdio transport for IDE integration
• Remote HTTP/SSE transport for Netlify or Vercel
• Example tools with Zod schema validation
• Uses @modelcontextprotocol/sdk

Python (FastMCP)

MCP server using the FastMCP Python framework:

• Data-driven servers with sample query tools
• Local stdio support for IDE integration
• Optional remote deployment to FastMCP Cloud
• Includes test script and comprehensive README
• Uses fastmcp>=2.0.0

Generated Project Structure

JavaScript Projects

my-mcp-server/
├── src/
│   ├── index.js       # Local MCP server (stdio transport)
│   └── tools.js       # Tool definitions with Zod schemas
├── netlify/
│   └── functions/
│       └── api.js     # Netlify Function (SSE transport)
├── data/              # Optional data files for your tools
├── package.json
├── netlify.toml
└── README.md

• src/tools.js — Define your MCP tools here (name, description, schema, handler)
• src/index.js — Stdio server for local IDE integration
• netlify/functions/api.js — SSE server for remote deployment
• data/ — Store static data files your tools need to access

FastMCP Projects

my-fastmcp-server/
├── server.py          # FastMCP server with sample tools
├── requirements.txt   # Python dependencies
├── test_server.py     # Validation script
├── data/
│   └── sample.json    # Sample data file
└── README.md          # Deployment guide

• server.py — FastMCP server with example data query tools
• test_server.py — Test script to validate functionality
• data/sample.json — Example data structure
• requirements.txt — Python dependencies (fastmcp>=2.0.0)

IDE Configuration

Add to your IDE's MCP settings:

JavaScript (Local):

{
  "mcpServers": {
    "my-server": {
      "command": "node",
      "args": ["/absolute/path/to/my-mcp-server/src/index.js"]
    }
  }
}

Python (Local):

{
  "mcpServers": {
    "my-fastmcp-server": {
      "command": "python",
      "args": ["/absolute/path/to/my-fastmcp-server/server.py"]
    }
  }
}

Remote (after Netlify deployment):

{
  "mcpServers": {
    "my-server": {
      "command": "npx",
      "args": ["mcp-remote@next", "https://my-server.netlify.app/mcp"]
    }
  }
}

Remote (after FastMCP Cloud deployment):

{
  "mcpServers": {
    "my-fastmcp-server": {
      "type": "http",
      "url": "https://your-project-name.fastmcp.app/mcp"
    }
  }
}

Deploy to Netlify

Via GitHub

1. Push your project to GitHub
2. Go to Netlify → "Add new site" → "Import an existing project"
3. Connect your repository — Netlify auto-detects settings from netlify.toml

Via CLI

npm install -g netlify-cli
netlify login
netlify deploy --prod

Endpoints

Once deployed:

• SSE: https://your-site.netlify.app/.netlify/functions/api/sse
• Message: https://your-site.netlify.app/.netlify/functions/api/message

Then rebuild: npm run build

Deploy to FastMCP Cloud

For Python projects generated with the FastMCP option:

1. Push your project to GitHub
2. Visit fastmcp.cloud and sign in with GitHub
3. Create new project from your repository
4. Configure:
   • Name: Choose a unique name (creates URL)
   • Entrypoint: server.py:mcp
   • Authentication: Public or organization-only
5. Deploy (automatic)
6. Connect at: https://your-project-name.fastmcp.app/mcp

FastMCP Cloud automatically redeploys on every push to main.

Requirements

• Node.js >= 18.0.0 (for JavaScript projects)
• Python >= 3.8 (for FastMCP projects)

License

MIT