Quick answer: State-machine-based agent framework — production-grade multi-step agents.
LangGraph is an open-source, state-machine-based agent framework built by LangChain for developing production-grade multi-step agents. It provides a structured approach to building agentic workflows by modeling agent behavior as explicit state transitions and decision points, rather than relying on implicit control flow. LangGraph enables developers to define complex agent interactions with clear state management, error handling, and human-in-the-loop capabilities. The framework is particularly useful for scenarios where reproducibility, observability, and fine-grained control over agent execution are critical. By combining graph-based workflows with LangChain's LLM integration ecosystem, LangGraph allows teams to move beyond simple prompt-response patterns into sophisticated multi-turn reasoning systems with guaranteed state consistency.