A zero-trust, AI-powered healthcare ecosystem where patients own their data, doctors work with AI intelligence, and medicines are verified on-chain.
Built for Hacktropica — Privacy-first healthcare for the modern world.
Healthcare data is fragmented, siloed, and controlled by institutions — not patients. Counterfeit medicines kill over 1 million people annually. Doctors are buried under cognitive overload from unstructured records. Emergency patients can't be identified without ID cards.
InfinityCare is a unified, multi-role healthcare platform built on zero-trust principles:
- 🔐 Patients own and control who sees their medical records
- 🤖 AI extracts structured clinical intelligence from raw prescriptions
- 🧬 Biometric Face ID identifies unconscious patients in emergencies
- ⛓️ Blockchain verifies medicine authenticity end-to-end
- 📡 Contactless vitals via Presage camera-based monitoring (no wearables)
┌─────────────────────────────────────────────────────────────────┐
│ FRONTEND (Next.js 16) │
│ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ │
│ │ Patient │ │ Doctor │ │ Hospital │ │ Nurse │ │
│ │Dashboard │ │Dashboard │ │Dashboard │ │Dashboard │ │
│ └────┬─────┘ └────┬─────┘ └────┬─────┘ └────┬─────┘ │
│ ┌────┴─────┐ ┌────┴─────┐ │
│ │ Pharmacy │ │ Vendor │ │
│ │Dashboard │ │Dashboard │ │
│ └──────────┘ └──────────┘ │
├─────────────────────────────────────────────────────────────────┤
│ API LAYER (Next.js Routes) │
│ /api/ai/summarize · /api/ai/predict · /api/auth/* │
│ /api/hospital/* · /api/patient/* · /api/doctor/* │
│ /api/my-documents · /api/verify-medicine · /api/pharmacy/* │
├─────────────────────────────────────────────────────────────────┤
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────────────┐ │
│ │ PostgreSQL │ │ Cloudinary │ │ Algorand Blockchain │ │
│ │ (Prisma) │ │ (File CDN) │ │ (Medicine Ledger) │ │
│ └──────────────┘ └──────────────┘ └──────────────────────┘ │
├─────────────────────────────────────────────────────────────────┤
│ PYTHON MICROSERVICE (Flask) │
│ DeepFace Biometric Engine · Face ID Matching │
├─────────────────────────────────────────────────────────────────┤
│ EXTERNAL AI SERVICES │
│ Gemini 2.5 Flash · Ollama · Disease Prediction API │
│ Presage SmartSpectra · Tesseract.js (Browser OCR) │
└─────────────────────────────────────────────────────────────────┘
| Feature | Description |
|---|---|
| Upload Documents | Encrypt and store lab reports, prescriptions, insurance docs in your personal vault (Cloudinary CDN) |
| Access Control | Grant/revoke per-file access to hospitals — toggle at any time |
| AI Prescription Scan | Upload a prescription photo → Tesseract.js OCR → Gemini AI structured summary (symptoms, medicines, dosage) |
| Disease Prediction | Extracted symptoms are fed to an ML model that returns ranked disease probabilities |
| Save & Share Summaries | Save AI-generated summaries and selectively share them with your treating doctor |
| Medicine Verification | Scan a medicine QR to verify authenticity against the Algorand blockchain |
| Real-time Notifications | Get instant alerts when a hospital accesses your documents |
| View Prescriptions | See digital prescriptions written by registered doctors in the network |
| Feature | Description |
|---|---|
| Type Prescription | Create structured digital prescriptions that go directly to the patient's vault |
| Prescription QR | Generate scannable QR codes for pharmacies to instantly verify prescriptions |
| AI Patient Summaries | View Gemini-generated clinical summaries shared by patients — structured, not raw |
| Previous Prescriptions | Review your authored prescription history for continuity of care |
| Feature | Description |
|---|---|
| Search Patient | Find patients and request decryption access to their shared medical records |
| Access Logs | View immutable audit trails of every document access for HIPAA compliance |
| Write Prescription | Author prescriptions directly into a patient's decentralized vault |
| Feature | Description |
|---|---|
| Emergency Face ID Match | Capture a face via camera → DeepFace CNN matches against the patient database instantly |
| Biometric Identification | Identify unconscious/unresponsive patients without ID cards in emergency situations |
| Instant Record Access | Once identified, immediately access the patient's permitted medical vault |
| Feature | Description |
|---|---|
| Scan Prescription QR | Scan doctor-issued QR codes to receive structured, authorized dispensing instructions |
| Medicine Details | Look up medicine records and safe dosage instructions |
| Verify Authenticity | Scan a medicine QR to check if it's genuine on the blockchain ledger |
| Dispensing History | Track every prescription dispensed for compliance and safety |
| Feature | Description |
|---|---|
| Register Medicine | Add medicine batches with manufacturer, expiry, and quantity details onto Algorand |
| Cryptographic Hash | Generate tamper-proof SHA-256 hashes from medicine payload for mathematical verification |
| Verification QR | Create scannable QR codes containing blockchain-backed authenticity proofs |
| Authenticity Registry | View all registered batches and their current verification status on the public ledger |
InfinityCare uses Google Gemini 2.5 Flash as the primary AI engine for transforming raw medical text into structured clinical intelligence.
Prescription Image → Tesseract.js OCR → Gemini 2.5 Flash → Structured JSON
├─ patientOverview
├─ symptoms[]
├─ medicines[]
├─ dosage
└─ notes
Resilient fallback chain ensures zero downtime:
- Gemini 2.5 Flash — Primary (structured JSON output, 1-3s response)
- Ollama Cloud (LLaMA) — Secondary (OpenAI-compatible)
- Custom Healthcare AI — Tertiary (self-hosted ML backend)
- Degraded Response — Graceful fallback (raw OCR text with warning)
A Flask microservice running DeepFace (VGG-Face CNN) enables contactless biometric identification:
- Patient face images are synced from PostgreSQL to a local cache
- Nurse captures a photo → DeepFace runs multi-layer CNN cosine similarity matching
- Returns patient ID + match confidence in under 2 seconds
- Zero wearable/hardware dependency — any webcam or phone camera works
Extracted symptoms from Gemini are forwarded to a separate ML API that returns ranked disease probabilities with confidence scores — giving patients and doctors a preliminary diagnostic signal.
Presage Technologies' SmartSpectra SDK extends InfinityCare's camera-first infrastructure for contactless vital sign monitoring:
| Vital | Method |
|---|---|
| Heart Rate (BPM) | Remote photoplethysmography (rPPG) |
| Heart Rate Variability (HRV) | Autonomic nervous system analysis |
| Breathing Rate | Respiratory cycle detection |
| Blood Pressure (Relative) | Pulse waveform analysis |
| Apnea Detection | Breathing pause identification |
Key integration points:
- Nurse Station — Extends the existing DeepFace camera flow to simultaneously capture vitals alongside patient identification (one scan = identity + vitals)
- Patient Self-Monitoring — At-home vitals via phone camera, stored in the encrypted medical vault
- Doctor AI Summaries — Presage vitals feed into the Gemini summary pipeline
InfinityCare uses the Algorand blockchain for medicine supply chain integrity:
- Vendors register medicine batches → cryptographic hash is generated and stored on-chain
- Pharmacies scan QR codes → hash is verified against the immutable ledger
- Patients can independently verify any medicine's authenticity
- Verification statuses:
GENUINE·NOT_FOUND·EXPIRED·TAMPERED·SUSPICIOUS
| Layer | Technology |
|---|---|
| Frontend | Next.js 16, React 19, TypeScript 5 |
| Styling | Tailwind CSS 4, Framer Motion, Glassmorphism UI |
| Database | PostgreSQL (Supabase), Prisma ORM 7 |
| Auth | Better Auth (session-based, role-gated) |
| File Storage | Cloudinary CDN |
| AI/LLM | Google Gemini 2.5 Flash, Ollama Cloud |
| OCR | Tesseract.js 7 (browser-side) |
| Biometrics | DeepFace (VGG-Face CNN), Flask/Python |
| Vitals | Presage SmartSpectra SDK |
| Blockchain | Algorand (algosdk) |
| QR Codes | qrcode + html5-qrcode |
| Forms | React Hook Form + Zod validation |
| Deployment | Vercel (frontend), Render (Python service) |
10 Prisma models powering the ecosystem:
User ──┬── Document ──── AccessLog
├── Prescription │
├── Notification │
├── Medicine AISummary
├── Session
└── Account
Verification
Roles: PATIENT · DOCTOR · HOSPITAL · PHARMACY · VENDOR
- Node.js 18+
- Python 3.9+ (for DeepFace service)
- PostgreSQL database (or Supabase)
git clone https://github.com/your-repo/infinitycare.git
cd infinitycare
npm installCreate a .env file in the project root:
# Database
DATABASE_URL="postgresql://user:password@host:5432/dbname"
# Auth
BETTER_AUTH_SECRET="your-auth-secret"
BETTER_AUTH_URL="http://localhost:3000"
# AI Services
GEMINI_API_KEY="your-gemini-api-key"
OLLAMA_API_KEY="your-ollama-key" # Optional fallback
OLLAMA_BASE_URL="https://ollama.com/v1" # Optional
OLLAMA_MODEL="llama3.2" # Optional
# File Storage
CLOUDINARY_CLOUD_NAME="your-cloud-name"
CLOUDINARY_API_KEY="your-api-key"
CLOUDINARY_API_SECRET="your-api-secret"
# Blockchain
NEXT_PUBLIC_ALGORAND_APP_ID="your-app-id"
# Python Service
NEXT_PUBLIC_FACE_MATCH_URL="http://localhost:5000"
# Disease Prediction
PREDICTION_API_BASE_URL="https://your-prediction-api.com"npx prisma generate
npx prisma db pushnpm run devcd src/python_service
python -m venv venv
.\venv\Scripts\activate # Windows
pip install -r requirements.txt
python app.pyThe app will be available at http://localhost:3000
healix-app/
├── prisma/
│ └── schema.prisma # 10 models, 5 roles, 4 enums
├── src/
│ ├── app/
│ │ ├── api/ # 14 API route groups
│ │ │ ├── ai/ # Gemini summarize + Disease predict
│ │ │ ├── auth/ # Better Auth endpoints
│ │ │ ├── doctor/ # Prescription & summary APIs
│ │ │ ├── hospital/ # Patient search & access logs
│ │ │ ├── patient/ # Notifications, save summaries
│ │ │ ├── pharmacy/ # QR scan & dispensing
│ │ │ └── ...
│ │ ├── patient/ # Patient dashboard page
│ │ ├── doctor/ # Doctor dashboard page
│ │ ├── hospital/ # Hospital dashboard page
│ │ ├── nurse/ # Nurse dashboard page
│ │ ├── pharmacy/ # Pharmacy dashboard page
│ │ ├── vendor/ # Vendor dashboard page
│ │ ├── verify/[hash]/ # Public medicine verification
│ │ └── page.tsx # Landing page
│ ├── components/
│ │ ├── patient/ # 7 modals (Upload, Docs, Verify, etc.)
│ │ ├── doctor/ # 3 modals (AI Summaries, QR, History)
│ │ ├── hospital/ # 3 modals (Search, Logs, Prescribe)
│ │ ├── nurse/ # 1 modal (Face ID Match)
│ │ ├── pharmacy/ # 3 modals (Scan QR, Medicine, History)
│ │ ├── vendor/ # 4 modals (Register, Hash, QR, Registry)
│ │ ├── auth/ # FaceCapture component
│ │ ├── layout/ # Shared dashboard layout
│ │ └── ui/ # Shadcn UI primitives
│ ├── lib/
│ │ ├── auth.ts # Better Auth server config
│ │ ├── auth-client.ts # Client-side auth
│ │ └── services/
│ │ └── analysisService.ts # AI summary + prediction client
│ └── python_service/
│ ├── app.py # Flask + DeepFace biometric engine
│ ├── requirements.txt
│ └── Procfile # Render deployment
└── package.json
| Service | Platform |
|---|---|
| Frontend | Vercel |
| Python/DeepFace | Render |
| Database | Supabase (PostgreSQL) |
| File CDN | Cloudinary |
Built with ❤️ for Hacktropica
InfinityCare — Because your health data should belong to you.