OpenAI Checkout Sample App

A full-stack e-commerce checkout application demonstrating the integration of OpenAI Apps SDK with Model Context Protocol (MCP) servers. This project consists of a Python backend MCP server and a Node.js/TypeScript frontend UI application.

📋 Project Overview

This application showcases how to build intelligent e-commerce experiences using:

• Model Context Protocol (MCP) – For LLM-server communication
• OpenAI Apps SDK – For rich UI components and widgets
• Conversational checkout – Natural language product search and purchase flow

📁 Project Structure

checkout-sample-app/
├── ch-openai-checkout-app/      # Backend MCP Server (Python)
└── ch-openai-checkout-ui/       # Frontend UI Application (Node.js/TypeScript)

1. ch-openai-checkout-app (Backend MCP Server)

Language: Python
Framework: FastMCP + Uvicorn
Port: 8080

Description: A Python-based MCP server that implements e-commerce tools for:

• Product search and catalog browsing
• Shopping cart management (add_to_cart, update quantities)
• Checkout session creation
• Payment processing and order completion

The server maintains shopping cart state across conversation turns using _meta["widgetSessionId"] to keep the UI widget in sync with backend state.

Key Features:

• SSE (Server-Sent Events) endpoint for MCP protocol
• Structured JSON responses with embedded UI metadata
• Shopping cart state management
• Checkout session handling

Files:

• shopping_cart_python/main.py – Main application entry point
• shopping_cart_python/models.py – Data models for products, carts, and sessions
• shopping_cart_python/requirements.txt – Python dependencies
• shopping_cart_python/Dockerfile – Container configuration (Python 3.12)
• product_data.json – Sample product catalog
• checkout_session_data.json – Checkout session templates
• checkout_completion_data.json – Order completion data

2. ch-openai-checkout-ui (Frontend UI Application)

Language: TypeScript/React
Framework: Vite + Tailwind CSS
Port: 8088

Description: A modern web-based UI application that renders the shopping cart widget and interfaces with the Python MCP server. Built with Vite for fast development and production builds.

Key Features:

• Shopping cart widget with real-time state sync
• Responsive design using Tailwind CSS
• Dynamic widget rendering from OpenAI Apps SDK
• Product display and checkout interface
• Health check actuator endpoint

Files:

• src/shopping-cart/ – Shopping cart widget components
• src/use-widget-state.ts – State management hook for widgets
• src/use-openai-global.ts – OpenAI SDK integration
• vite.config.mts – Vite build configuration
• package.json – Node.js dependencies
• pnpm-lock.yaml – Dependency lock file
• Dockerfile – Container configuration (Node 20 Alpine)

---

🚀 Getting Started

Prerequisites

• For Python Backend:

  • Python 3.12+
  • pip or poetry
  • Docker (optional, for containerized setup)

• For Node.js Frontend:

  • Node.js 20+
  • pnpm (recommended) or npm/yarn
  • Docker (optional, for containerized setup)

Setup Instructions

Option 1: Run Services Locally

Backend Setup (Python):

cd ch-openai-checkout-app

# Create virtual environment (optional but recommended)
python3.12 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r shopping_cart_python/requirements.txt

# Run the MCP server
python shopping_cart_python/main.py
# Server runs on http://localhost:8080

Frontend Setup (Node.js):

cd ch-openai-checkout-ui

# Install dependencies
pnpm install
# or: npm install

# Development server (hot reload)
pnpm run dev
# or: npm run dev
# Accessible at http://localhost:8088

# Build for production
pnpm run build
# or: npm run build

# Preview production build
pnpm serve

Option 3: Run with Docker

Build Backend:

cd ch-openai-checkout-app
docker build -f shopping_cart_python/Dockerfile -t checkout-app-backend .
docker run -p 8080:8080 checkout-app-backend

Build Frontend:

cd ch-openai-checkout-ui
docker build -f Dockerfile -t checkout-app-ui .
docker run -p 8088:8088 checkout-app-ui

---

🔄 How the Application Works

Conversation Flow

1. User Request → "I'd like to find a gift card"
2. MCP Tool Call → product_search tool is triggered
3. Product Results → Search results returned with product catalog
4. Add to Cart → User clicks add button in the shopping cart widget
5. State Sync → Widget state updates via _meta["widgetSessionId"]
6. Checkout → User clicks checkout to create a checkout session
7. Payment → Card capture screen appears for tokenization
8. Order Complete → Order confirmation is displayed

State Management

• Backend State: Cart maintained on the Python server
• Widget State: Synchronized via window.openai.widgetState
• Session Tracking: widgetSessionId keeps widget and server aligned across conversation turns
• Persistence: Checkout session ID enables multi-turn order tracking

---

📡 API Endpoints

Backend MCP Server (Port 8080)

• GET /mcp – Server-Sent Events (SSE) connection for MCP protocol
• POST /mcp/messages?sessionId={sessionId} – Process MCP tool calls
• GET /actuator/health – Health check endpoint

Frontend (Port 8088)

• / – Main shopping cart widget interface
• /actuator/health – Health check endpoint
• /assets/* – Static assets (CSS, JavaScript)

---

🛠️ Development

Backend Development

cd ch-openai-checkout-app

# Run with auto-reload
uvicorn shopping_cart_python.main:app --reload --host 0.0.0.0 --port 8080

Frontend Development

cd ch-openai-checkout-ui

# Start dev server with hot reload
pnpm run dev

When making changes to either service, follow conventions:

• Python: PEP 8 style guide, type hints recommended
• TypeScript: ESLint and Prettier configured, run pnpm run lint

---

📦 Key Dependencies

Backend (Python):

• FastMCP – MCP server framework
• Uvicorn – ASGI server
• Pydantic – Data validation
• Requests – HTTP client

Frontend (Node.js):

• React – UI framework
• TypeScript – Type safety
• Tailwind CSS – Utility-first CSS
• Vite – Build tool
• Pnpm – Package manager

---

📝 Environment Variables

Backend (.env in ch-openai-checkout-app)

APP_VERSION=1.0.0
HOST=0.0.0.0
PORT=8080

Frontend (.env in ch-openai-checkout-ui)

VITE_API_URL=http://localhost:8088
VITE_PORT=8088

---

🚢 Deployment

Docker Setup

Both services include Dockerfiles optimized for production:

• Backend: python:3.12-slim – Lightweight Python image
• Frontend: node:20-alpine – Minimal Node.js image

Kubernetes

For Kubernetes deployments, adapt the Docker images and expose services via Ingress.

---

📚 Additional Resources

• Model Context Protocol Documentation
• OpenAI Apps SDK
• FastMCP Documentation
• Vite Documentation

---

📄 License

MIT License – See LICENSE file for details

---

🤝 Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

---

📞 Support

For issues, questions, or suggestions, please open a GitHub issue or contact the maintainers.