SDK Apps Collection

A curated gallery of interactive applications developed with the OpenAI Apps SDK and Model Context Protocol.

Architecture • Projects • Getting Started • Repository

---

Overview

This repository demonstrates advanced integration patterns between Large Language Models (LLMs) and interactive user interfaces. By leveraging the Model Context Protocol (MCP), these applications allow ChatGPT to render dynamic widgets, visualize data in real-time, and perform complex actions beyond simple text generation.

The collection serves as a reference implementation for developers building the next generation of AI-native applications.

Projects

Project	Description	Key Technologies	Status
Real Estate Map	Interactive property search engine with dynamic map visualization and filtering capabilities.	Google Maps API, SSE, Zod, MCP	🟢 Active
Medical Appointment	Doctor scheduling system with availability checking, booking, and ambiguity handling.	Supabase, STDIO, Zod, MCP	🟢 Active

Architecture

All applications in this monorepo share a unified architectural pattern designed for scalability and maintainability:

• Protocol: Model Context Protocol (MCP) for standardized LLM communication.
• Transport: Server-Sent Events (SSE) over HTTP for persistent, bi-directional data flow.
• Backend: Node.js/TypeScript services implementing tool definitions and business logic.
• Frontend: Lightweight, zero-dependency HTML/JS widgets injected directly into the chat interface.

Getting Started

Follow these steps to set up the environment and run the demo applications locally.

Prerequisites

Requirement	Version	Purpose
Node.js	v18.0.0+	Javascript Runtime
npm	v9.0.0+	Package Manager
Google Maps Key	Active	Required for Real Estate App

Installation

1. Clone the repository:

   git clone https://github.com/tostechbr/sdk-apps-openai.git
   cd sdk-apps-openai

2. Navigate to the target application directory (e.g., Real Estate):

   cd apps/real-estate

3. Install dependencies:

   npm install

4. Configure environment variables (create .env file):

   cp .env.example .env
   # Edit .env with your specific API keys

Development Workflow

We recommend using the MCP Inspector for local development and testing of tool definitions before integration with ChatGPT.

# Start the development server
npm run dev

# In a separate terminal, launch the Inspector
npx @modelcontextprotocol/inspector sse http://localhost:8787/mcp

Contributing

We welcome contributions to expand this collection of use cases. Please ensure all new modules follow the established directory structure and include comprehensive documentation.

License

This project is licensed under the MIT License - see the LICENSE file for details.