AI-powered review analytics system. Ingests noisy, multilingual e-commerce reviews from simulated platform pages via API, processes them through a full backend intelligence pipeline, and delivers feature-level sentiment analysis, graph-based issue classification, trend detection, and prioritized recommendations through a React + enterprise dashboard.
Protected by an employee authentication gate — the login page is the first thing rendered on app load.
| File | Read when building... |
|---|---|
| docs/ARCHITECTURE.md | System overview, layers, data flow |
| docs/AUTH.md | Authentication — login page, JWT flow, credentials |
| docs/STACK.md | Tech stack, folder structure, environment |
| docs/DATA.md | Review schema, seed data, simulation pages |
| docs/API.md | All backend endpoints + request/response contracts |
| docs/PIPELINE.md | Backend processing stages 1–7 |
| docs/INSIGHTS.md | Insight types, health score, alerts, recommendations, PDF |
| docs/UI.md | React frontend, dashboard panels, Demo Center, animations |
| docs/PROGRESS.md | ✅ Build checklist — update after every session |
# Backend
cd server && npm install && npm run dev
# Frontend (main app)
cd client && npm install && npm run dev
# Simulation pages (static, open directly in browser)
open simulation/amazon.htmlCreate server/.env:
SARVAM_API_KEY=your-key
GEMINI_API_KEY=your-key
JWT_SECRET=your-secret-key-here
PORT=5000
Simulation pages post reviews to
http://localhost:5000/api/reviews/ingest. Main app frontend connects to the same backend.
| Field | Value |
|---|---|
| Employee ID | npd570 |
| Password | notre570 |
The Employee ID is shown as placeholder text in the login field. Wrong credentials show an "Authentication Failed" message.