Scannova

AI Market Research Agent for E-Commerce Product Discovery in Singapore

Live Demo · Report Bug · Request Feature

Screenshot

About

Scannova is an AI-powered market research agent that helps Singapore e-commerce sellers discover trending products, analyze pricing, and identify import opportunities. It searches across trusted regional platforms (Lazada, Shopee, Amazon SG, Qoo10) using live Google Search grounding and returns ranked, exportable results.

Key Features

• AI-Powered Search — Gemini 2.0 Flash with Google Search grounding for live product data
• Smart Fallback — Automatic OpenAI GPT-4o-mini fallback when Gemini is unavailable
• Trusted Sites — Lazada, Shopee, Amazon SG, Qoo10, and more
• Price Filtering — Set min/max price ranges in SGD
• Product Ranking — Scored by popularity (sales, reviews, demand signals)
• CSV Export — Download results for downstream analysis
• Premium UI — Animated, responsive design with live status indicators

Tech Stack

Layer	Technology
Frontend	React 19 + Vite 8
Styling	Vanilla CSS + Inter font
Backend	Express 5 (Node.js)
AI (Primary)	Google Gemini 2.0 Flash + Search
AI (Fallback)	OpenAI GPT-4o-mini
Deployment	GitHub Pages (frontend) + GitHub Actions

Architecture

┌──────────────┐     ┌──────────────────┐     ┌─────────────────────┐
│   React UI   │────▶│   Express API    │────▶│ Gemini 2.0 (Primary)│
│  (Vite SPA)  │     │  /api/research   │     │   + Google Search   │
└──────────────┘     └────────┬─────────┘     └─────────────────────┘
                              │                          │
                              │ (fallback on failure)    │
                              ▼                          ▼
                     ┌─────────────────────┐
                     │  OpenAI GPT-4o-mini │
                     └─────────────────────┘

Note: The GitHub Pages deployment hosts the React frontend only. To use the AI research features, run the Express backend locally (or deploy it to a Node host such as Render, Railway, or Fly.io) and point the frontend at it.

Project Structure

scannova/
├── .github/workflows/deploy.yml   # GitHub Pages CI/CD
├── server/
│   ├── index.js                   # Express API server
│   └── services/
│       ├── gemini.js              # Gemini API integration
│       ├── openai.js              # OpenAI fallback
│       └── researcher.js          # Orchestrator
├── src/
│   ├── App.jsx                    # Main app component
│   ├── index.css                  # Design system
│   ├── main.jsx                   # React entry
│   └── components/
│       ├── Header.jsx
│       ├── SearchForm.jsx
│       ├── ProductCard.jsx
│       ├── ResultsPanel.jsx
│       └── LoadingState.jsx
├── .env.example                   # API key template
├── package.json
└── vite.config.js

Getting Started

Prerequisites

• Node.js 20+
• A Google Gemini API key
• (Optional) An OpenAI API key for fallback

1. Clone & install

git clone https://github.com/alfredang/scannova.git
cd scannova
npm install

2. Configure API keys

cp .env.example .env

Edit .env:

GEMINI_API_KEY=your_gemini_key_here
OPENAI_API_KEY=your_openai_key_here

3. Run locally

Terminal 1 — Backend API:

npm run server

Terminal 2 — Frontend:

npm run dev

Open http://localhost:5173.

Deployment

GitHub Pages (frontend)

This repo ships with a GitHub Actions workflow at .github/workflows/deploy.yml that builds the Vite SPA and publishes it to GitHub Pages on every push to main.

Live site: https://alfredang.github.io/scannova/

Backend

The Express server in server/ is a standalone Node.js app. Deploy it to any Node host (Render, Railway, Fly.io, a VM, etc.) and update the frontend to point at its URL.

Contributing

1. Fork the repo
2. Create a feature branch (git checkout -b feat/your-feature)
3. Commit your changes
4. Open a Pull Request

Issues and feature requests are welcome on the issues page.

Developed By

Tertiary Infotech Academy Pte. Ltd.

Acknowledgements

• Google Gemini for the primary research model and Search grounding
• OpenAI for the fallback model
• Vite and React for the frontend toolchain

---

If you find this useful, please consider starring the repo to support the project.