LangGraph
langchain-ai.github.io/langgraph
A low-level framework from LangChain for building stateful, multi-agent AI applications as explicit graphs, with controllable execution flow, human-in-the-loop support, and built-in persistence.
Why it is useful
Solves the control problem with LLM agents — instead of hoping the agent reasons correctly through a chain, you define the possible states and transitions explicitly as a graph. This makes agent behavior predictable, debuggable, and resumable across sessions. The human-in-the-loop support lets you pause execution at any node for approval before continuing, which is essential for production agentic workflows.