An AI-powered pizza ordering assistant built with FastAPI, LangChain, Google Gemini, and a Vite + React frontend.
PizzaBot supports conversational ordering, policy/FAQ lookup, and flexible vector search.
cd backend
cp .env.example .env # set GOOGLE_API_KEY (and other keys if needed)
pip install -r requirements.txt
uvicorn app.main:app --reload
cd frontend
npm install
npm run dev
curl -X POST http://localhost:8000/api/ingest \
-H 'Content-Type: application/json' \
-d '[{"id":"faq1","text":"Delivery within 5km. Cash or card."}]'
📂 Project Structure
pizzabot/
├── backend/ # FastAPI + LangChain services
│ ├── app/ # routers, graph, config, schemas
│ ├── .env.example # environment template
│ └── requirements.txt
├── frontend/ # Vite + React client
│ ├── src/ # components, pages, API calls
│ └── package.json
└── README.md
🏗️ Architecture
Routers → Graph → LLM / Vector → Config / Schemas
Loose coupling with dependency injection for easier testing & swapping components
Pydantic structured output ensures consistent contracts for UI
Vector DB abstraction: swap Chroma for another database without touching routers or UI
LangSmith tracing auto-enabled if LANGSMITH_API_KEY is set
Gemini-compatible via langchain-google-genai (ChatGoogleGenerativeAI)
🌐 Tech Stack
Backend: FastAPI, LangChain, ChromaDB, Google Gemini
Frontend: React, Vite, Bootstrap
Infra: dotenv, Pydantic, Uvicorn
🔑 Environment Variables
Create .env inside backend/ (see .env.example):
# Google Gemini
GOOGLE_API_KEY=your_api_key_here
# LangSmith (optional)
LANGSMITH_API_KEY=your_langsmith_key_here
# Backend Config
ENV=dev
HOST=0.0.0.0
PORT=8000
VECTORDB_DIR=./data/chroma
FRONTEND_ORIGIN= http://localhost:5173
🚀 Roadmap
Save orders to database
User authentication & profiles
Payment gateway integration
Multi-LLM backend support (OpenAI, Anthropic, etc.)
Deployment with Docker
🤝 Contributing
Contributions are welcome!
Fork the repo, create a feature branch, and open a Pull Request.
📜 License
MIT License © 2025 PizzaBot Authors