🩺 GlukoBuddy — MediQ

AI-Based Early Diabetes Risk Monitoring System for Children

🏆 2nd Place — Hacknovate 2K26 | Smart Healthcare Domain

---

Overview

Most diabetes tools give parents a number and leave them confused. GlukoBuddy gives them context they can actually act on.

Instead of:

"Risk Score: 28%"

GlukoBuddy explains:

"Recent glucose fluctuations combined with reduced sleep and energy levels may indicate early diabetic risk patterns. Monitoring sugar intake and maintaining insulin schedules is recommended."

Built in 30 hours by a team of 4 during Hacknovate 2K26.

---

Features

Feature	Description
🩺 AI Risk Prediction	Custom weighted scoring engine across glucose, BMI, sleep, stress, heredity
🤖 Explainable AI Insights	LLM-generated health narratives via Ollama (llama3.2)
📊 30-Day Trend Analysis	Monthly report summarization with risk trend detection
🐶 KidsBuddy	Gamified questionnaire designed for children
🍫 NutriFinder	Food label scanner for nutrition/sugar tracking
📲 Guardian Alerts	Real-time SMS alerts to parents via Twilio on high risk
👨‍⚕️ Doctor Dashboard	Separate role-based view for medical professionals
💬 AI Chatbot	WebSocket-based healthcare chatbot with session memory & summarization
📄 PDF Reports	Downloadable longitudinal health records
🔍 OCR Lab Reports	Upload PDFs or images — text extracted via Tesseract.js
🌐 Multilingual	Marathi locale support built in
⏰ Insulin Reminders	Scheduled medication alerts

---

Tech Stack

Frontend

• React + Vite + Tailwind CSS
• Context API (Auth, Theme, Language)
• Socket.io client for real-time chat

Backend

• Node.js + Express.js
• MongoDB + Mongoose
• Socket.io for WebSocket chat
• JWT + bcrypt for auth
• Joi for request validation
• Multer for file uploads
• Tesseract.js + pdf-parse for OCR
• Twilio for SMS alerts
• node-cron for scheduled reminders

AI Layer

• Ollama (local LLM inference)
• Model: llama3.2:1b
• Used for: risk analysis, chatbot replies, chat summarization, monthly report summaries

---

Architecture

React Client
    │
    ├── WebSocket ──────► Socket.io (Chat + Live Monitoring)
    │
    └── REST API ───────► Express Server
                              │
                 ┌────────────┼────────────┐
                 │            │            │
              MongoDB      Ollama      Twilio
            (Health DB)  (LLM Layer)  (SMS Alerts)
                              │
                         Tesseract.js
                         (OCR Engine)

---

Local Setup

⚠️ This project runs a local LLM via Ollama and is not deployed.

Prerequisites

• Node.js v18+
• MongoDB (local or Atlas)
• Ollama installed with llama3.2:1b pulled

ollama pull llama3.2:1b

Installation

# Clone
git clone https://github.com/shaikhsiddique/MediQ
cd MediQ

# Backend
cd Backend
npm install
cp .env.example .env
# Fill in your values
npm run dev

# Frontend
cd ../Frontend
npm install
npm run dev

Environment Variables

PORT=4000
MONGODB_URI=mongodb://127.0.0.1:27017/mediq
JWT_SECRET=your_secret
CLIENT_URL=http://localhost:5173
API_PUBLIC_URL=http://localhost:4000

# Ollama
OLLAMA_BASE_URL=http://localhost:11434
OLLAMA_MODEL=llama3.2:1b

# Twilio (optional — for SMS alerts)
TWILIO_ACCOUNT_SID=
TWILIO_AUTH_TOKEN=
TWILIO_PHONE_NUMBER=
PARENT_PHONE_NUMBER=

---

Team

Name	Profile
Siddique Shaikh	@shaikhsiddique
Omkar Ghorpade	LinkedIn
Sai Jadhav	LinkedIn
Yash Jagdale	LinkedIn

---

Hackathon

• Event: Hacknovate 2K26
• Domain: Smart Healthcare
• Duration: 30 hours
• Result: 🥈 2nd Place

---

License

MIT