Kenko AI is a modern preventive health intelligence platform. It combines a Next.js 16 frontend, local Python-based risk inference, and Gemini-powered natural language insights to help users understand and act on their health data.

• Local AI Inference: Secure, fast risk prediction using local Python models (no cloud upload required)
• Gemini-Powered Insights: Human-friendly, context-aware explanations for every result
• Animated UI: GSAP and Tailwind for a beautiful, responsive experience
• Downloadable Reports: Export your results as JSON for sharing or further analysis
• Context-Aware Inputs: Glucose context (fasting/random), BMI calculator, and more

1. Clone the repo:

   git clone <your-repo-url>
   cd kenko

2. Install dependencies:

   pnpm install

3. Configure Gemini (optional): Create a .env.local file:

   GEMINI_API_KEY=your_gemini_api_key
   GEMINI_MODEL=gemini-3-flash-preview

   If not set, the app will still run and provide fallback insights.
4. Run the app:

   pnpm dev

   Visit http://localhost:3000

Frontend:

• Next.js 16 (app router, TypeScript, Tailwind CSS)
• GSAP for animations
• VitalsForm for structured health input

Backend/API:

• /api/predict route spawns a Python process for local model inference
• Python risk engine normalizes, predicts, and explains risk
• Gemini API (if configured) generates natural language insights

ML Models:

• Trained offline using OpenML datasets (see kenko-model)
• Models and training reports stored in kenko-model/models

kenko/
├── app/
│   ├── api/predict/route.ts      # API handler (Node <-> Python <-> Gemini)
│   ├── components/               # UI components (VitalsForm, Results, etc.)
│   └── ...
├── public/
├── styles/
├── .env.local                    # Gemini API keys (not committed)
└── README.md

1. Enter your age, BMI, glucose, blood pressure, activity, diet, sleep, fatigue, and smoking status.
2. Click Predict Trajectory.
3. View:
   • Metabolic, cardiovascular, and nutritional risk scores
   • Top contributing factors
   • Gemini-powered plain-language summary
   • Downloadable JSON report

Pull requests are welcome! For major changes, please open an issue first to discuss what you’d like to change.

This project is for hackathon and research use. For production or clinical use, consult a medical professional and review all code and models for safety and compliance.