AI Chat with LangGraph, Google Gemini, and SSE

This project provides a streaming AI chat API using FastAPI, LangGraph, and Google Generative AI (Gemini) with tool integration (Tavily Search) and Server-Sent Events (SSE).
It is designed to integrate seamlessly with a Next.js frontend.

---

🚀 Features

• 🧠 Streaming AI responses using SSE for real-time chat experience.
• 🔍 Tool integration: Supports dynamic tool calls (Tavily Search).
• 💬 Chat memory: Remembers previous conversations using LangGraph checkpointing.
• 💾 Checkpointing: Maintains chat context with UUID-based checkpoints for each user/session.
• 🌐 CORS enabled: Easily integrates with frontend apps.
• ⚡ Next.js ready: Built for seamless streaming with a Next.js frontend.

---

🧩 Tech Stack

• Backend: FastAPI, LangGraph, LangChain Google Generative AI
• AI Model: ChatGoogleGenerativeAI with gemini-2.5-flash
• Tools: Tavily Search Results
• Memory: LangGraph MemorySaver for checkpointing and chat persistence
• Frontend: Next.js with SSE support
• Environment Management: python-dotenv