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

vercel-mcp-python/
├── api/
│   └── index.py          # Main Vercel function
├── src/
│   └── mcp_server.py     # Your MCP server logic
├── client-app/           # Interactive MCP client
│   ├── mcp_client.py     # Rich client application
│   ├── requirements.txt  # Client dependencies
│   ├── setup.py          # Setup script
│   ├── README.md         # Client documentation
│   └── run_client.bat    # Windows launcher
├── requirements.txt       # Server dependencies
├── vercel.json           # Vercel configuration
└── README.md

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:

Install Node.js (if not already installed):
- Download from nodejs.org
Install Vercel CLI globally:
```
npm install -g vercel
```
Verify installation:
```
vercel --version
```

If you encounter PATH issues:

# Find npm global directory
npm config get prefix

# Add to PATH (add this to your ~/.bashrc)
export PATH=$PATH:$(npm bin -g)
source ~/.bashrc

Alternative methods:

# Using npx (no global installation)
npx vercel

# Using yarn
yarn global add vercel

Setup

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

Install dependencies:
```
pip install -r requirements.txt
```
Login to Vercel:
```
vercel login
```
Deploy to Vercel:
```
vercel --prod
```

Local Development

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

vercel dev

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 vercel dev:

Solution 1: Deploy directly (Recommended)

vercel --prod

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

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

# Activate virtual environment first
.\venv\Scripts\Activate.ps1  # Windows PowerShell
# or
source venv/Scripts/activate  # Windows Git Bash

# Test MCP server functionality
python -c "
import sys, os
sys.path.append('src')
from mcp_server import mcp
import json

# Test echo tool
request = {
    'jsonrpc': '2.0',
    'id': 1,
    'method': 'tools/call',
    'params': {'name': 'echo', 'arguments': {'message': 'Hello from venv!'}}
}
response = mcp.handle_request(request)
print(json.dumps(response, indent=2))
"

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

You can test it by:

Opening the URL in your browser
Using a tool like Postman or curl
Connecting with an MCP client

Solution 4: Run as Administrator (if needed)

Close your terminal
Right-click on Git Bash/PowerShell and select "Run as administrator"
Navigate back to your project: cd /d/repos/vercel-mcp-python
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:

# Navigate to client directory
cd client-app

# Setup (first time only)
python setup.py

# Configure environment (optional)
cp .env.example .env
# Edit .env to customize server URL and settings

# Run the client
python mcp_client.py

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