FastAPI Backend for AI-Powered Document Intelligence

FastAPI application with Supabase, LangChain, and AI integrations for the MemexLLM AI research assistant.

Frontend Repo • Features • Tech Stack • Quick Start • Deployment

The MemexLLM backend provides a powerful API for document processing, AI-powered chat, and content generation. It leverages modern AI frameworks to transform documents into interactive knowledge bases.

Built for production with async processing, streaming responses, and enterprise-grade security.

The backend follows a service-oriented architecture with a focus on asynchronous processing and reliable document ingestion.

• Document Processing: PDF, TXT, DOCX, audio, YouTube, and web content extraction
• AI Chat: Contextual Q&A with source citations and streaming responses
• Content Generation: Create podcasts, quizzes, flashcards, mindmaps, and more
• Vector Search: Semantic search using Pinecone embeddings
• Authentication: Supabase Auth with JWT tokens
• Real-time: Server-Sent Events (SSE) for streaming responses

• Python 3.11+
• PostgreSQL database (local or Supabase)
• API keys for AI providers

cd backend
uv sync

Copy .env.example to .env and configure:

# Supabase
SUPABASE_URL=https://your-project.supabase.co
SUPABASE_KEY=your-service-role-key

# Database
DATABASE_URL=postgresql://user:pass@localhost:5432/memexllm

# AI Providers
OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...
GOOGLE_API_KEY=...

# Pinecone
PINECONE_API_KEY=...
PINECONE_INDEX=memexllm

uv run uvicorn src.main:app --reload

API docs available at http://localhost:8000/docs

backend/
├── src/
│   ├── api/              # API routes
│   │   └── v1/
│   │       ├── notebooks.py
│   │       ├── chat.py
│   │       ├── sources.py
│   │       └── ...
│   ├── core/             # Core config
│   │   ├── config.py
│   │   ├── security.py
│   │   └── supabase.py
│   ├── models/           # SQLAlchemy models
│   ├── services/         # Business logic
│   │   ├── ai/
│   │   ├── chat/
│   │   ├── documents/
│   │   └── storage/
│   └── main.py           # App entry point
├── migrations/            # Alembic migrations
├── tests/                # Test files
└── alembic.ini           # Alembic config

docker-compose up -d

See docs/COOLIFY_DEPLOYMENT.md for detailed deployment instructions.

docker-compose -f docker-compose.prod.yml up -d

See http://localhost:8000/docs for full API documentation.

1. Follow the existing code style (ruff)
2. Write tests for new features
3. Update documentation

MIT License