Product Listing Application

This project is now implemented as a Next.js App Router application.

Run locally

1. Install dependencies:
   • npm install
2. Copy environment variables:
   • copy .env.example .env.local (Windows)
3. Set API credentials in .env.local:
   • BRIGHTDATA_MCP_URL (recommended for OpenAI MCP, e.g. https://mcp.brightdata.com/mcp?...)
   • BRIGHTDATA_MCP_SSE_URL (legacy/direct SSE fallback)
   • BRIGHTDATA_API_KEY (optional if token is embedded in MCP URL)
   • API_TOKEN (supported alias for Bright Data key, matches MCP config)
   • GROUPS / TOOLS (optional source tool filter, matches MCP config)
   • OPENAI_API_KEY (required for OpenAI MCP integration path)
   • OPENAI_MODEL (optional, default gpt-4.1-mini)
   • ACONTEXT_API_KEY
   • ACTIONBOOK_API_KEY
4. Optional MCP stdio config (same values used by this app when it launches npx @brightdata/mcp):

{
  "mcpServers": {
    "Bright Data": {
      "command": "npx",
      "args": ["@brightdata/mcp"],
      "env": {
        "API_TOKEN": "YOUR_TOKEN",
        "GROUPS": "advanced_scraping",
        "TOOLS": "web_data_amazon_product,web_data_walmart_product,web_data_ebay_product,web_data_etsy_products,web_data_bestbuy_products,web_data_google_shopping"
      }
    }
  }
}

5. Start dev server:
   • npm run dev

Swagger Testing

1. Open Swagger UI:
   • /api/docs
2. Open raw OpenAPI spec:
   • /api/openapi
3. Use Try it out in Swagger UI to test:
   • POST /api/compare/start
   • GET /api/compare/{requestId}
   • GET /api/saved?sessionId=...
   • POST /api/saved/toggle
   • POST /api/saved/alert

Main capabilities

1. Responsive dashboard and product pages with mobile sidebar behavior.
2. Saved deals with shared bookmark state and price alerts.
3. Compare API + polling endpoints.
4. Bright Data integration via OpenAI Responses MCP first (type: "mcp" to Bright Data), then direct MCP (SSE/stdio) and @brightdata/sdk fallback, with source normalization and ranking.
5. Acontext + Actionbook adapter hooks for context and playbook usage.