MedRx - Medical Report Analyzer

MedRx is a full-stack MVP that lets users upload medical reports (JPG/PNG/PDF), extract report text via OCR, detect key health parameters, and generate simple explanations in English or Hinglish.

Tech Stack

• Frontend: React + Vite + Tailwind CSS
• Backend: Express + Next.js (custom server integration)
• Database: MongoDB (Mongoose)
• Auth: JWT + bcrypt
• OCR/PDF Parsing: Tesseract.js + pdf-parse

Folder Structure

PROJECT FOR SEM 6/
├── frontend/
│   ├── src/
│   │   ├── components/
│   │   ├── pages/
│   │   ├── services/
│   │   └── utils/
│   └── ...
└── backend/
    ├── next-client/pages/index.js
    ├── src/
    │   ├── config/
    │   ├── controllers/
    │   ├── middleware/
    │   ├── models/
    │   ├── routes/
    │   └── services/
    ├── server.js
    └── .env.example

Step-by-Step Setup

1) Backend Setup

cd backend
cp .env.example .env

Update .env values:

• MONGO_URI (your MongoDB connection string)
• JWT_SECRET (any strong random secret)

Run backend:

npm install
npm run dev

Backend runs at http://localhost:5000.

2) Frontend Setup

cd frontend
npm install
npm run dev

Frontend runs at http://localhost:5173.

App Flow

1. Splash page auto-redirects after 1.5 seconds.
2. Home page has "Start Now".
3. Portal has split layout:
   • Left: Login/Signup (MongoDB + JWT)
   • Right: Upload report (works with or without login)
4. File is analyzed:
   • OCR text extraction
   • Medical value detection (Hemoglobin, Blood Sugar, Cholesterol)
   • Normal/Low/High comparison
   • Natural language explanation
5. Results page shows extracted text, parameter cards, and explanation with language toggle.

API Endpoints

• POST /api/auth/signup
• POST /api/auth/login
• POST /api/analyze (multipart file field: report)
• GET /api/history (optional logged-in history)
• GET /api/health

Notes for Production

• Replace local storage token strategy with HTTP-only cookies.
• Add request validation (e.g., Zod/Joi).
• Integrate queue/worker for OCR at scale.
• Use object storage (S3/GCS) instead of local temp uploads.