LangGraph overview

LangGraph is a low-level orchestration framework and runtime for long-running, stateful agents. You define workflows as graphs of shared state, nodes, and edges, and can mix deterministic steps with LLM-driven steps in the same graph. LangGraph focuses on orchestration capabilities such as durable execution, streaming, and human-in-the-loop. It does not prescribe prompts or agent architecture. Before you start, learn the building blocks agents use, especially models and tools. Docs examples often use LangChain for those integrations, but LangGraph does not require LangChain. If you want a higher-level agent loop first, use Deep Agents or LangChain agents. Use LangGraph when you need explicit control over state and control flow.

Show How LangChain products fit together

Deep Agents is an agent harness: planning, subagents, filesystem tools, and context management on top of LangGraph.
LangChain is the agent framework: abstractions and integrations for models, tools, and agent loops.
LangGraph is the orchestration runtime: durable execution, streaming, human-in-the-loop, and persistence.
LangSmith is the platform for tracing, evaluation, prompts, and deployment across frameworks.
LangSmith Engine detects issues in LangGraph agent traces and proposes fixes. You can open a pull request with the proposed fix from the Engine tab.
LangSmith Fleet is the no-code agent builder for templates, integrations, and routine automation.

For a comparison of the open source stack, see Frameworks, runtimes, and harnesses.

Install

pip install -U langgraph

uv add langgraph

Create a minimal graph:

from langgraph.graph import StateGraph, MessagesState, START, END

def mock_llm(state: MessagesState):
    return {"messages": [{"role": "ai", "content": "hello world"}]}

graph = StateGraph(MessagesState)
graph.add_node(mock_llm)
graph.add_edge(START, "mock_llm")
graph.add_edge("mock_llm", END)
graph = graph.compile()

graph.invoke({"messages": [{"role": "user", "content": "hi!"}]})

For a full calculator agent, see the LangGraph quickstart.

Capabilities

LangGraph provides the runtime primitives for stateful, long-running workflows. It does not abstract prompts or force a fixed agent architecture.

Deterministic and agentic steps: Combine hand-coded nodes with model-driven nodes in one graph.
Persistence: Checkpoint state and resume after failures or pauses.
Human-in-the-loop: Pause execution, inspect or edit state, then continue.
Memory: Short-term thread state and longer-term memory across sessions.
Streaming: Stream tokens, state updates, and other events as the graph runs.

Integrations, deployment, and observability

LangGraph works on its own. However, you can also pair it with other LangCHain products:

LangChain

Integrations and agent abstractions built on LangGraph.

LangSmith Deployment

Deploy and scale stateful agents, with visual prototyping in Studio.

LangSmith Observability

Trace requests, evaluate outputs, and monitor deployments.

Get started

Capabilities

Patterns

Build with the Graph API

Deployment

Frontend

Advanced

Install

Capabilities

Integrations, deployment, and observability

LangChain

LangSmith Deployment

LangSmith Observability

See also

​ Install

​Capabilities

​Integrations, deployment, and observability

LangChain

LangSmith Deployment

LangSmith Observability

​See also

Install

Capabilities

Integrations, deployment, and observability

See also