APU - Privacy-Preserving Medical AI

AI-powered medical diagnosis system with complete privacy using Fully Homomorphic Encryption (FHE) and decentralized AI.

🏆 Features ERC-7857 (Agentic ID): Tokenize AI agents as NFTs with encrypted metadata

Architecture

Dual-Network Design

• Ethereum Sepolia: FHE encrypted medical data (Zama fhEVM)
• 0G Network: Decentralized storage and TEE-verified AI inference

Technology Stack

• Smart Contracts: Solidity 0.8.27 with Zama FHE + ERC-7857 (Agentic ID)
• Backend: Node.js with Drizzle ORM
• Frontend: Next.js 16 with Privy authentication
• Storage: 0G Storage with AES-256-GCM encryption
• AI: 0G Compute with TEE attestation

ERC-7857 Implementation

APU implements the official Agentic ID standard (ERC-7857) for tokenizing AI agents:

• 📄 Contracts: APUAgenticID.sol, IERC7857.sol
• 🔑 Features: NFT-based AI ownership, encrypted metadata, TEE-verified transfers
• 📊 Documentation: See ERC7857_IMPLEMENTATION.md for complete technical details
• 🎯 Use Case: Medical AI agents as tradeable, privacy-preserving NFTs

Only 3-4 of 21 projects in 0G Bridge by AKINDO Wave 3 have implemented this standard.

Getting Started

Prerequisites

• Node.js >= 20
• PostgreSQL
• Private keys for deployment

Installation

# Install dependencies
npm install

# Setup environment
cp .env.example .env
# Edit .env with your configuration

# Setup database
npm run db:push --workspace=packages/backend

Development

# Start frontend
npm run dev

# Compile contracts
npm run contracts:compile

# Run backend
npm run backend:dev

Deployment

# Deploy to Sepolia
npm run contracts:deploy:sepolia

# Setup 0G integration
npm run contracts:deploy:og

Project Structure

apu/
├── packages/
│   ├── contracts/          # Hardhat smart contracts
│   │   ├── contracts/      # Solidity contracts
│   │   └── scripts/        # Deployment scripts
│   └── backend/            # Backend services
│       ├── src/
│       │   ├── db/         # Drizzle ORM schema
│       │   └── services/   # 0G integration
│       └── drizzle.config.ts
└── app/                    # Next.js frontend
    └── src/
        ├── app/            # App router pages
        ├── components/     # React components
        └── lib/            # Utilities

Key Features

• Privacy-First: FHE ensures data never decrypted on-chain
• Decentralized AI: 0G Compute with TEE attestation
• Cross-Chain: Dual-network architecture
• User-Friendly: Email login with embedded wallets (Privy)
• Production-Ready: Gas sponsorship, monitoring, compliance

Live Deployments

Production Environments

• Frontend: https://apu-frontend.onrender.com
• Backend API: https://apu-backend-7a8z.onrender.com

Smart Contracts (Sepolia)

• MedicalDataRegistry: 0x2819Cf40a952748014C56f393e1ffd16f4a377ff

  • Etherscan

• APUAgenticID (ERC-7857): 0xE619B84c5837E43512cA219f0bffa6c9A290Ba99

  • Etherscan

0G Network Integration

• Storage: 0G Storage Testnet (Galileo)
• Compute: 0G Compute TEE with ECDSA verification
• Network: https://evmrpc-testnet.0g.ai

Quick Start

Run the automated setup script:

./setup.sh

Or manually:

# 1. Install dependencies
npm install

# 2. Setup environment
cp .env.example .env
# Edit .env with your values

# 3. Create database
createdb apu
npm run db:push --workspace=packages/backend

# 4. Compile contracts
npm run contracts:compile

# 5. Start development server
npm run dev

Testing

# Test 0G Storage integration
npm run test:storage --workspace=packages/backend

# Test 0G Compute integration
npm run test:compute --workspace=packages/backend

# Test contracts
npm run test --workspace=packages/contracts

# Test frontend
npm run test --workspace=app

Security

• All patient data encrypted with FHE (Zama)
• Medical history encrypted with AES-256-GCM
• AI diagnosis runs in TEE (Trusted Execution Environment)
• Doctor authorization enforced on-chain
• KMS Gateway for decryption control

Contributing

Contributions welcome! Please read the development guide first.

License

MIT