Mediloon AI-Driven Agentic Pharmacy System

🎯 Overview

An intelligent, autonomous pharmacy ordering system that uses AI agents to manage medicine ordering, predict refills, and automate procurement workflows. Features voice/text ordering, predictive analytics, and multi-agent automation.

✨ Key Features

1. Voice/Text Ordering Assistant

• 🎤 Natural language ordering via voice or text
• 🌍 Multi-language support (English, German, Arabic)
• 🧠 Intelligent medicine name parsing
• 💬 Multi-turn conversation capability
• ✅ Prescription validation
• 🛒 Automatic cart creation

2. Predictive Ordering Engine

• 📊 Learns from customer order history
• 🔮 Predicts medicine depletion dates
• 🔔 Auto-suggests refills via notifications
• 📈 Visual prediction timelines
• 🎯 Personalized refill recommendations

3. Multi-Agent System (LangChain/LangGraph)

• Ordering Agent: Assists customers during ordering
• Forecast Agent: Predicts refill and stock needs
• Procurement Agent: Automates purchase orders
• Safety Agent: Validates prescriptions and checks interactions

4. Workflow Automation (MCP Integration)

• 📧 Email/SMS/WhatsApp notifications via Zapier
• 🔄 Distributor API calls via n8n
• 🔗 Webhook integration with CMS
• ⚡ Real-time inventory updates

🏗️ Architecture

mediloon-pharmacy/
├── client/                # React frontend (JavaScript/JSX)
│   ├── src/
│   │   ├── pages/        # Page components
│   │   ├── context/      # React contexts
│   │   └── App.jsx       # Main app
│   └── public/           # Static assets
├── server/               # Node.js Express backend
│   ├── controllers/     # Request handlers
│   ├── routes/          # API routes
│   ├── models/          # Database models
│   ├── middleware/      # Auth, validation
│   └── sockets/         # Real-time handlers
├── ai-backend/          # Python FastAPI AI server
│   ├── agents/          # LangChain agents
│   ├── api/             # REST API endpoints
│   ├── models/          # Database models
│   └── core/            # AI logic

🚀 Tech Stack

Frontend (Client)

• Framework: React 18 (Create React App)
• Language: JavaScript (JSX)
• Styling: CSS + inline styles
• UI Components: Custom components
• Voice: Web Speech API + Whisper
• State: React Context
• Real-time: Socket.io client
• Routing: React Router v6

Backend (Server)

• Framework: Node.js Express
• Language: JavaScript
• Database: MongoDB (Mongoose)
• Cache: Redis
• Authentication: JWT + bcrypt
• Real-time: Socket.io server
• Validation: express-validator

AI Backend

• Framework: FastAPI (Python 3.11+)
• AI/ML: LangChain, LangGraph, OpenAI GPT-4
• Database: MongoDB (Motor async driver)
• Cache: Redis
• Voice: OpenAI Whisper

📦 Installation

Prerequisites

• Node.js 18+ and npm
• Python 3.11+
• MongoDB 6.0+ (running locally or MongoDB Atlas)
• Redis 7.0+ (running locally)

1. Setup Databases

Install and start MongoDB and Redis locally:

MongoDB:

• Download from https://www.mongodb.com/try/download/community
• Start: mongod --dbpath <your-data-path>

Redis:

• Download from https://redis.io/download
• Start: redis-server

2. Backend Setup (AI FastAPI)

cd ai-backend

# Activate virtual environment
venv\Scripts\activate  # Windows
# or: source venv/bin/activate  # Mac/Linux

# Install dependencies
pip install -r requirements.txt

# Set environment variables
copy .env.example .env  # Windows
# or: cp .env.example .env  # Mac/Linux
# Edit .env with your OpenAI API key

# Start server
python main.py

Runs on: http://localhost:8000

3. Server Setup (Node.js Express)

cd server

# Install dependencies
npm install

# Set environment variables
copy .env.example .env  # Windows
# or: cp .env.example .env  # Mac/Linux
# Edit .env with your JWT_SECRET and database URLs

# Start server
npm run dev

Runs on: http://localhost:3000

4. Client Setup (React)

cd client

# Install dependencies
npm install

# Set environment variables
copy .env.example .env  # Windows
# or: cp .env.example .env  # Mac/Linux

# Start development server
npm start

Runs on: http://localhost:3001

5. Access Application

• Customer Interface: http://localhost:3001
• API Server: http://localhost:3000
• AI Backend API: http://localhost:8000/docs
• Health Checks:
  • Express: http://localhost:3000/health
  • FastAPI: http://localhost:8000/health

🔑 Environment Variables

AI Backend (.env)

# Database
MONGODB_URI=mongodb://localhost:27017/mediloon
REDIS_URL=redis://localhost:6379

# AI/ML
OPENAI_API_KEY=your_openai_api_key

# Security
JWT_SECRET=your_jwt_secret_key

Server (.env)

NODE_ENV=development
PORT=3000
MONGODB_URI=mongodb://localhost:27017/mediloon
REDIS_URL=redis://localhost:6379
JWT_SECRET=your_jwt_secret_key
FASTAPI_URL=http://localhost:8000
FASTAPI_API_KEY=your_internal_api_key
CORS_ORIGIN=http://localhost:3001

Client (.env)

REACT_APP_API_URL=http://localhost:3000
REACT_APP_SOCKET_URL=http://localhost:3000

🎮 Usage

Customer Ordering

1. Voice Ordering

   • Click the microphone button
   • Say: "I need 30 tablets of Aspirin 100mg"
   • Confirm the parsed order
   • Complete checkout

2. Text Ordering

   • Type in the chat: "I need metformin 500mg"
   • The AI assistant will guide you through options
   • Add to cart and checkout

3. Predictive Refills

   • View your medicine timeline
   • Get automatic refill notifications
   • One-click reorder

Admin Operations

1. Inventory Management

   • Monitor stock levels
   • View predictive alerts
   • Approve auto-procurement

2. Agent Monitoring

   • View agent activity logs
   • Monitor conversation quality
   • Review safety alerts

3. Analytics

   • Order patterns
   • Prediction accuracy
   • Customer satisfaction

🤖 Multi-Agent System

Agent Workflow

graph TD
    A[Customer Request] --> B[Ordering Agent]
    B --> C{Valid Order?}
    C -->|Yes| D[Safety Agent]
    C -->|No| B
    D --> E{Prescription Required?}
    E -->|Yes| F[Request Prescription]
    E -->|No| G[Forecast Agent]
    F --> G
    G --> H{Stock Available?}
    H -->|Yes| I[Add to Cart]
    H -->|No| J[Procurement Agent]
    J --> K[Order from Supplier]
    K --> I

Loading

Agent Descriptions

Ordering Agent

• Handles natural language conversations
• Parses medicine names and quantities
• Manages multi-turn dialogues
• Confirms alternatives and variants

Forecast Agent

• Analyzes order history
• Predicts depletion dates
• Calculates refill timelines
• Triggers proactive notifications

Procurement Agent

• Monitors inventory levels
• Creates purchase orders automatically
• Integrates with supplier APIs via n8n
• Optimizes order quantities

Safety Agent

• Validates prescription requirements
• Checks drug interactions
• Verifies dosage appropriateness
• Flags potential safety issues

🔧 API Endpoints

Express Server (Port 3000)

Authentication

POST   /api/auth/register       Register new user
POST   /api/auth/login          Login user
POST   /api/auth/logout         Logout user
POST   /api/auth/refresh        Refresh token
GET    /api/auth/me             Get current user
PATCH  /api/auth/profile        Update profile

Orders

GET    /api/orders              Get user orders
GET    /api/orders/:id          Get order by ID
POST   /api/orders/process      Process order (proxy to AI)

Medicines

GET    /api/medicines           List medicines
GET    /api/medicines/:id       Get medicine details

FastAPI AI Backend (Port 8000)

AI Ordering

POST   /api/v1/orders/text      Process text order
POST   /api/v1/orders/voice     Process voice order
POST   /api/v1/orders/confirm   Confirm order
GET    /api/v1/orders/user/:id  Get user orders

Predictions

GET    /api/v1/predictions/user/:id   Get user predictions
GET    /api/v1/medicines              List medicines

🧪 Testing

# Backend tests
cd backend
pyServer tests
cd server
npm test

# AI Backend tests
cd ai-backend
pytest tests/ -v --cov

# Client tests
cd client
Docker deployment will be added in future updates. For now, follow manual deployment:

1. Set up MongoDB and Redis on your server
2. Deploy AI Backend (FastAPI) - see [ai-backend/README.md](ai-backend/README.md)
3. Deploy Server (Express) - see [server/README.md](server/README.md)
4. Build and deploy Client (React) - see [client/README.md](client/README.md)

Configure environment variables for production and use a process manager like PM2 for Node.js service
### Manual Deployment

See [DEPLOYMENT.md](./DEPLOYMENT.md) for detailed deployment instructions.

## 📊 Monitoring

- **Express Health**: `GET http://localhost:3000/health`
- **FastAPI Health**: `GET http://localhost:8000/health`
- **Socket.io Status**: Check browser console for connection status

## 🔒 Security

- JWT-based authentication with bcrypt password hashing
- Role-based access control (RBAC)
- Socket.io authentication via JWT
- Express rate limiting
- Input validation with express-validator
- CORS configuration
- Environment variable protection

## 🌍 Multi-Language Support

Supported languages:
- 🇬🇧 English
- 🇩🇪 German (Deutsch)
- 🇸🇦 Arabic (العربية)

## 📝 License

MIT License

## 🆘 Support

- **Issues**: Create an issue on GitHub
- **Documentation**: See individual README files in server/, client/, ai-backend/

## 🎯 Roadmap

- ✅ Voice/text ordering with AI
- ✅ Multi-agent system (LangChain)
- ✅ Real-time updates via Socket.io
- ✅ JWT authentication
- ✅ React frontend with routing
- 🔄 Complete AI integration (frontend → AI backend)
- 🔄 Voice recording and Whisper integration
- 🔄 Admin dashboard
- 🔄 Analytics and reporting
- 🔄 Docker containerization
- 🔄 Production deployment

---

**Built with ❤️ for HackFusion Hackathon**

- [ ] Mobile app (React Native)
- [ ] Voice assistant integration (Alexa, Google Home)
- [ ] Blockchain for prescription verification
- [ ] AR medicine scanner
- [ ] Telemedicine integration
- [ ] Insurance claim automation

## 👥 Team

Built with ❤️ by the Mediloon team for HackFusion 2025

---

**Note**: This is a hackathon project demonstrating AI-driven pharmacy automation. For production use, ensure compliance with local pharmaceutical regulations and data privacy laws.