GeminiGuard Backend Docs

API Endpoints

Real-time Chat

WebSocket /chat/stream

Stream AI responses using Gemini API with RAG from FAISS vectors.

                        {
  "type": "user_message",
  "content": "How do I reset my password?"
}
                    

FAQ Upload

POST /faq/upload

Upload CSV/Markdown files to generate embeddings for RAG pipeline.

curl -X POST -F "file=@faqs.csv" http://localhost:8000/faq/upload

Escalation

POST /escalate

Create support tickets stored in PostgreSQL with optional SendGrid email.

                        {
  "issue": "Payment failed",
  "description": "Card declined despite sufficient funds",
  "customer_email": "user@example.com"
}
                    

Escalation List

GET /escalate/list

Retrieve all support tickets with filtering options.

                        [
  {
    "id": 1,
    "status": "open",
    "created_at": "2023-07-15T09:30:00Z",
    "customer_email": "user@example.com"
  }
]
                    

Setup Guide

Environment Variables

                        GEMINI_API_KEY=your_gemini_key
REDIS_URL=redis://default:password@localhost:6379
POSTGRES_URL=postgresql://user:password@localhost:5432/supportdb
SENDGRID_KEY=your_sendgrid_key
CORS_ORIGINS=https://your-frontend.vercel.app,http://localhost:3000
                    

Running with Docker

                        # Build the image
docker build -t geminiguard-backend .

# Run the container
docker run -p 8000:8000 --env-file .env geminiguard-backend
                    

Directory Structure

                        backend/
├── main.py                # FastAPI app entrypoint
├── routes/                # API endpoint definitions
│   ├── chat.py            # WebSocket chat endpoint
│   ├── faq.py            # FAQ upload endpoint
│   └── escalate.py       # Ticket endpoints
├── services/             # Business logic
│   ├── gemini_service.py # Gemini API calls
│   └── faiss_service.py # FAISS vector operations
├── models/               # Pydantic models
├── database/             # DB connections
├── requirements.txt      # Python dependencies
└── Dockerfile            # Container configuration
                    

Ready to Deploy?

This backend is production-ready for Railway deployment with all required endpoints.

Download Complete Code

GeminiGuard Support Backend