harmanjeet-kaur-3/smart-shop-ai

Purpose and Runtime Entry Points

SmartShop AI is a full-stack product comparison application: users search and compare products, receive AI recommendations, and manage authentication and wishlists. The React/Vite client starts at src/main.tsx and renders application routing/UI from src/App.tsx. The Node/Express API is entered through server.ts.

Frontend Application Shell

The client is organized as route-level experiences backed by reusable product and recommendation UI:

• Product discovery/comparison: src/pages/Home.tsx, src/pages/Search.tsx, src/components/ProductCard.tsx

• Authenticated user areas: src/pages/Dashboard.tsx, src/pages/History.tsx, src/pages/Wishlist.tsx

• Authentication and input/upload flows: src/pages/Auth.tsx, src/pages/Upload.tsx

• Shared navigation and AI presentation: src/components/Navbar.tsx, src/components/AIRecommendation.tsx

Client-to-API Boundary

src/services/api.ts centralizes browser calls to the Express backend, while src/types.ts defines frontend domain contracts. UI pages invoke this service rather than directly coupling to server routes. The Vite runtime is configured by vite.config.ts; index.html is its document entry point.

HTTP API and Access Control

The Express server divides core capabilities into route modules:

• Identity/session endpoints: server/routes/auth.ts

• Product lookup and search orchestration: server/routes/search.ts

• Saved-product management: server/routes/wishlist.ts

• Usage or product analytics: server/routes/analytics.ts

Authentication enforcement is a middleware boundary in server/middleware/auth.ts, separating protected user data flows from public API behavior.

Product Enrichment and AI Pipeline

Search requests can cross two service boundaries: product acquisition/scraping in server/services/scraper.ts, then AI-driven comparison, specification analysis, or recommendations in server/services/ai.ts. The AI service is expected to use the externally configured Groq API via GROQ_API_KEY, despite the README’s earlier Gemini wording. Results are returned through API routes for rendering by src/components/AIRecommendation.tsx and product-oriented pages.

Persistence and Infrastructure

The application uses a local JSON-backed data store, represented by data/db.json and accessed through server/db.ts. This is the persistence boundary for user-associated state such as wishlists and potentially history/analytics; it is not a separately provisioned database service. Runtime infrastructure is TypeScript across client and server, React/Vite in the browser build path, and Node.js/Express for the API process.