This article details testing a LangGraph-based AI agent's backend with Postgres. The agent manages a booking process, reading and writing data. Two persistence modes exist: in-memory for quick tests and Postgres for durable storage.

To test locally, set up a .env file with an OpenAI key and optionally a DATABASE_URL. Using Docker, you can run a Postgres container easily, ensuring the same connection string each time. The agent's state and bookings persist across sessions with Postgres, unlike in-memory mode.

The full source code is available on GitHub for cloning and testing.