---
title: ClientSphere RAG Backend
emoji: 🤖
colorFrom: blue
colorTo: purple
sdk: docker
sdk_version: "4.0.0"
python_version: "3.11"
app_file: app.py
pinned: false
---

# ClientSphere RAG Backend

FastAPI-based RAG (Retrieval-Augmented Generation) backend for ClientSphere AI customer support platform.

## Features

- 📚 Knowledge base management
- 🔍 Semantic search with embeddings
- 💬 AI-powered chat with citations
- 📊 Confidence scoring
- 🔒 Multi-tenant isolation
- 📈 Usage tracking and billing

## API Endpoints

- `GET /health/live` - Health check
- `GET /kb/stats` - Knowledge base statistics
- `POST /kb/upload` - Upload documents
- `POST /chat` - Chat with RAG
- `GET /kb/search` - Search knowledge base

## Environment Variables

Required:
- `GEMINI_API_KEY` - Google Gemini API key
- `ENV` - Set to `prod` for production
- `LLM_PROVIDER` - `gemini` or `openai`

Optional:
- `ALLOWED_ORIGINS` - CORS allowed origins (comma-separated)
- `JWT_SECRET` - JWT secret for authentication

## Documentation

See `README_HF_SPACES.md` for deployment details.