Definition
What is LangGraph?
Last updated
Definition
LangGraph is a Python and TypeScript framework for building stateful, multi-agent LLM applications as graphs of nodes with conditional edges and persistent state.
LangGraph builds on LangChain primitives but adds explicit state management, conditional routing, and human-in-the-loop interrupts. Nodes are functions; edges are routing decisions made by the graph or by an LLM. State persists across calls and can be checkpointed to disk or a database. It's the framework underneath most production LangChain agents that need retries, branching, or HITL approval. Glitch Grow's Ads Operator and Sales Agent are both LangGraph-based.
Why graphs and not chains
Linear chains can’t represent the loops, branches, and retries that real agents need. LangGraph’s graph model maps directly to the underlying control flow without forcing it into a sequence.
State persistence
LangGraph checkpoints state at each node, so an agent can be paused (waiting for HITL approval), resumed across processes, or replayed for debugging. This is the feature that turns a prototype into a production agent.