gemini-fullstack-langgraph-quickstart

Why we included this project

This repo is a working reference for anyone building a research-augmented conversational agent. It pairs a React/Vite frontend with a LangGraph backend that turns a plain question into a loop: generate search queries, query the web through the Google Search API, reflect on what's still missing, and refine until it can produce a cited answer. The project is laid out so you can follow the agent graph step by step, run the backend logic from the command line for quick tests, and then bring up the full stack with Docker Compose, including the Redis and Postgres pieces LangGraph needs for streaming and persistent state. Teams evaluating LangGraph for real products will find the deployment notes and project structure a reasonable starting point.