Enterprise-grade Compliance-First AI Content Generation & Validation POC with strict role separation and deterministic rule enforcement.

• Docker & Docker Compose
• API keys configured in .env file

docker-compose up -d

This will:

1. Start PostgreSQL database
2. Start FastAPI backend (runs seed script automatically)
3. Start React frontend

• Frontend: http://localhost:3000
• Backend API: http://localhost:8000
• API Docs: http://localhost:8000/docs

• Agent: agent_user
• Admin: admin_user
• Super Admin: super_admin_user

• /backend/app/models/ - SQLAlchemy ORM models
• /backend/app/providers/ - Cloud-swappable LLM & vector DB abstractions
• /backend/app/services/ - Business logic (rule management, content generation, compliance checking)
• /backend/app/api/ - FastAPI endpoints (agent, admin, super_admin)

• /frontend/src/components/ - React components by role
• /frontend/src/pages/ - Page components
• /frontend/src/services/ - API client
• /frontend/src/types/ - TypeScript definitions

✅ Rule-First Architecture - Rules override AI outputs
✅ Duplicate Detection - SQL exact match + Pinecone semantic similarity
✅ Immutable Versioning - All rule updates create new versions
✅ Token-Based Chunking - Preserves legal meaning in documents
✅ Audit Trail - Every action logged
✅ Cloud-Swappable - Provider abstraction for easy migration

• Backend: FastAPI, SQLAlchemy, PostgreSQL
• LLMs: Gemini / Groq (swappable)
• Vector DB: Pinecone (swappable to OpenSearch)
• Frontend: React, TypeScript, Vite
• Deployment: Docker Compose

cd backend
pip install -r requirements.txt
python seed_data.py
uvicorn app.main:app --reload

cd frontend
npm install
npm run dev

Visit http://localhost:8000/docs for interactive API documentation.

This is a POC. For production:

• Add proper authentication (OAuth/JWT)
• Implement rate limiting
• Add comprehensive logging
• Set up monitoring and alerting
• Configure CORS properly
• Use managed database services
• Implement proper secret management