fastapi-langgraph-agent-production-ready-template

Why we included this project

Moving from a LangGraph demo to something you can actually run in production usually means building a pile of supporting infrastructure yourself. This template does that work for you: it connects FastAPI to a stateful LangGraph agent with checkpointing, tool calling, and human-in-the-loop, then adds the pieces a live service needs, including mem0 long-term memory backed by pgvector, JWT auth with session management, rate limiting, structured logging, Langfuse tracing, and Prometheus metrics. A circular model fallback with retries and a timeout budget keeps one flaky provider from taking down the whole API. If you want a real starting point for an agent backend rather than assembling middleware by hand, this gives you a coherent base to build from.