Vercel MCP Python Server

A Model Context Protocol (MCP) server built with Python and FastMCP, designed to run on Vercel's serverless platform.

Project Structure

Features

This MCP server provides the following tools:

• echo: Echo back a provided message

• get_time: Get the current server time

• add_numbers: Add two numbers together

• get_weather_info: Get mock weather information for a location

And the following resources:

• config://server: Server configuration information

Prerequisites

Before setting up the project, you'll need to install the Vercel CLI:

Installing Vercel CLI

For Git Bash on Windows:

1. Install Node.js (if not already installed):

   • Download from nodejs.org

2. Install Vercel CLI globally:

   npm install -g vercel

3. Verify installation:

   vercel --version

If you encounter PATH issues:

Alternative methods:

Setup

1. Create and activate virtual environment (Recommended):

   # Create virtual environment python -m venv venv # Activate virtual environment # On Windows PowerShell: .\venv\Scripts\Activate.ps1 # On Windows Git Bash: source venv/Scripts/activate # On macOS/Linux: source venv/bin/activate

2. Install dependencies:

   pip install -r requirements.txt

3. Login to Vercel:

   vercel login

4. Deploy to Vercel:

   vercel --prod

Local Development

To test locally, you can use Vercel's development server:

Troubleshooting Windows Issues

Note: Local development with vercel dev may have issues on Windows due to runtime initialization errors. This is a known limitation and doesn't affect production deployment.

If you encounter issues with

Solution 1: Deploy directly (Recommended)

Your server will be available at the provided Vercel URL and works perfectly in production.

Solution 2: Test locally with Python (in virtual environment)

Solution 3: Use the deployed version Your server will be available at your Vercel domain after deployment.

You can test it by:

1. Opening the URL in your browser

2. Using a tool like Postman or curl

3. Connecting with an MCP client

Solution 4: Run as Administrator (if needed)

1. Close your terminal

2. Right-click on Git Bash/PowerShell and select "Run as administrator"

3. Navigate back to your project: cd /d/repos/vercel-mcp-python

4. Try vercel dev again

API Endpoints

• GET /: Returns server information and status

• POST /: Handles MCP protocol requests

• OPTIONS /: Handles CORS preflight requests

Dependencies

• fastmcp>=0.15.0: FastMCP framework for building MCP servers

• uvicorn>=0.24.0: ASGI server for Python web applications

• python-json-logger>=2.0.0: JSON logging for Python applications

Configuration

The server is configured through vercel.json with:

• Python runtime using @vercel/python

• 30-second maximum execution time

• CORS enabled for cross-origin requests

• Automatic routing to the main handler

Usage

Once deployed, your MCP server will be available at your Vercel domain. You can connect to it using any MCP-compatible client.

Using the Included Client App

A rich, interactive client application is included in the client-app/ directory:

The client provides:

• 🔌 Connection testing

• 🔧 Interactive tool calling

• 📚 Resource management

• 🧪 Automated testing of all tools

• 🎨 Beautiful console interface

See client-app/README.md for detailed usage instructions.

Additional Resources

• Vercel MCP Documentation - Official Vercel documentation for Model Context Protocol

• MCP Servers Repository - Explore available MCP servers

• AI SDK Documentation - Use the AI SDK to initialize MCP clients

License

MIT