Find cheaper alternatives (dupes) for products across Indian e-commerce platforms.

A ChatGPT App powered by the Model Context Protocol (MCP)

DesiDupe helps Indian users find "dupes" — cheaper alternatives for products they love. Whether you find an expensive item on Myntra or a pricey gadget on Amazon, DesiDupe scans across Flipkart, Amazon.in, Meesho, Myntra, Ajio, and Tata Cliq to find similar products at lower prices.

• 🔍 Product Search - Natural language search across all platforms
• 🏷️ Find Dupes - Paste any product URL and get cheaper alternatives
• 💰 Price Comparison - Compare the same product across different sites
• 📸 Image Analysis - Describe a product you saw and find similar items
• 🔥 Best Deals - Get the hottest discounts across categories

User: Find cheaper alternatives for this dress: https://www.myntra.com/dresses/zara/zara-floral-midi/XXX111

DesiDupe: Finding Dupes for: Zara Floral Print Midi Dress
Original Price: ₹3,990 on Myntra

Found 3 Cheaper Alternatives:

🏆 Best Dupe: Floral Print Maxi Dress
   Price: ₹399 (Save ₹3,591)
   Platform: Meesho
   Rating: 3.7⭐ (34,567 reviews)

User: Compare prices for iPhone 15 128GB across sites

DesiDupe: Price Comparison: iPhone 15 (128GB)

| Platform  | Price      | Status       |
|-----------|------------|--------------|
| Amazon    | ₹73,999 🏷️ | ✅ In Stock  |
| Flipkart  | ₹74,999    | ✅ In Stock  |
| Tata Cliq | ₹76,999    | ✅ In Stock  |

Best Price: ₹73,999 on Amazon
You Save: ₹3,000 vs Tata Cliq

User: Find wireless earbuds under ₹1000

DesiDupe: Found 3 products

1. pTron Bassbuds Duo - ₹599 on Meesho (70% off)
2. Noise Buds VS104 - ₹1,099 on Flipkart (63% off)

• Node.js 18+
• npm or yarn

# Clone or navigate to the project
cd desidupe

# Install dependencies
npm install

# Build TypeScript
npm run build

# Run the MCP server
npm start

desidupe/
├── src/
│   ├── mcp/
│   │   ├── server.ts         # MCP server entry point
│   │   ├── tools.ts          # Tool definitions & schemas
│   │   └── handlers/
│   │       ├── searchHandler.ts   # search_product, get_deals
│   │       ├── dupesHandler.ts    # find_dupes, analyze_image
│   │       └── compareHandler.ts  # compare_prices
│   ├── services/
│   │   ├── scraper.ts        # URL parsing & product fetching
│   │   ├── search.ts         # Search logic
│   │   └── similarity.ts     # Dupe finding & price comparison
│   ├── data/
│   │   └── mockProducts.ts   # Mock product database
│   └── types/
│       └── index.ts          # TypeScript types
├── app-manifest.json         # OpenAI Apps SDK manifest
├── package.json
├── tsconfig.json
└── README.md

Search for products across all platforms with natural language support.

{
  "name": "search_product",
  "arguments": {
    "query": "budget earbuds under Rs 1500",
    "platforms": ["flipkart", "amazon"],
    "sortBy": "price_asc"
  }
}

Find cheaper alternatives for a product.

{
  "name": "find_dupes",
  "arguments": {
    "productUrl": "https://www.myntra.com/...",
    "maxResults": 5
  }
}

Compare prices across platforms.

{
  "name": "compare_prices",
  "arguments": {
    "productName": "iPhone 15 128GB"
  }
}

Get best deals by category.

{
  "name": "get_deals",
  "arguments": {
    "category": "electronics",
    "maxResults": 10
  }
}

Find products similar to what's in an image.

{
  "name": "analyze_image",
  "arguments": {
    "imageDescription": "red floral summer dress",
    "maxPrice": 2000
  }
}

npx @modelcontextprotocol/inspector node dist/mcp/server.js

# Build and run
npm run build
node dist/mcp/server.js

# In another terminal, send JSON-RPC requests via stdin

Add to your ChatGPT MCP configuration:

{
  "mcpServers": {
    "desidupe": {
      "command": "node",
      "args": ["/path/to/desidupe/dist/mcp/server.js"]
    }
  }
}

Run the server locally and configure ChatGPT to connect via stdio.

Deploy as an HTTP endpoint using the MCP HTTP transport.

Submit to the OpenAI Apps marketplace when available.

• MCP server structure
• Mock product data
• Basic tools (search, dupes, compare)
• App manifest

• Flipkart API integration
• Amazon affiliate API
• Web scraping for other sites
• Product image matching

• Redis caching
• Rate limiting
• Analytics dashboard
• User preferences

• More platforms (Nykaa, Croma, etc.)
• Price drop alerts
• Wishlist feature
• Browser extension

Contributions welcome! Key areas:

1. Real API integrations
2. Better product matching algorithms
3. More product categories
4. UI improvements

MIT

Built with ❤️ for Indian shoppers