LangGraph
OrchestrationMIT
LangGraph is an orchestration framework and runtime for agents that run over a long time and keep state. A workflow is a graph of shared state, nodes and edges, and it can mix deterministic steps with model-driven ones. Its product page describes it as the framework to build agents with low-level control. The Python package installs with pip, a separate JavaScript library exists, and the open-source library can be used without LangChain. Optional deployment goes through the LangSmith Agent Server on paid plans.
The documentation describes durable execution through checkpointed state, pausing for human review, short-term and long-term memory, and streaming of tokens and state updates.

Pinout
- MCP server: not checked yet
- HTTP API: not checked yet
- command line: not checked yet
- Python SDK: supported, verified
- TypeScript SDK: supported, verified
- OpenAPI description: not checked yet
| Interface | Status | Source we read | Read on |
|---|---|---|---|
| MCP server | Not checked yet | — | — |
| HTTP API | Not checked yet | — | — |
| Command line | Not checked yet | — | — |
| Python SDK | Supported, verified | Documentation overview, Install (pip install -U langgraph) | |
| TypeScript SDK | Supported, verified | README, link to the LangGraph.js library | |
| OpenAPI description | Not checked yet | — | — |
Facts
| Pricing | Hosting runs through LangSmith Agent Server: Developer $0 a seat without Agent Server, Plus $39 a seat a month with one small serverless Agent Server included, Enterprise priced on request. LangChain pricing page, read . |
|---|---|
| License | MIT; the product page calls it an MIT-licensed open-source library, free to use. LangGraph product page, read . |
| Python install | pip install -U langgraph (or uv add langgraph). Documentation overview, Install, read . |
| JavaScript | A separate library, LangGraph.js, is referenced for JavaScript/TypeScript. README in the repository, read . |
| Core features | Durable execution, human-in-the-loop, memory, streaming. Documentation overview, orchestration capabilities, read . |
| Workflow model | Graphs of shared state, nodes and edges; deterministic and LLM steps can be mixed. Documentation overview, read . |
| Deployment models | Cloud (managed in LangSmith), hybrid (self-hosted data plane), fully self-hosted. LangGraph product page, deployment section, read . |
| Deployment requirement | Agent Server deployment is available in the LangSmith Plus and Enterprise plans. LangGraph product page, deployment section, read . |
| Latest release | langgraph 1.2.22. GitHub Releases page, read . |
| Independence | Can be used without LangChain. Documentation overview, read . |
| Last commit | . GitHub API, latest commit on the default branch of langchain-ai/langgraph, read . |
| Frameworks | LangChain (Documentation overview, integrations, read 2026-11-16). |
Call example
graph.invoke({"messages": [{"role": "user", "content": "hi!"}]})Limits
- The Agent Server deployment options are available in the LangSmith Plus and Enterprise plans; no price is read here.
- The documentation overview covers Python only; JavaScript is a separate library.
Strengths
- The library is MIT-licensed and installs with a single pip command.
- It offers durable execution, human-in-the-loop pauses, memory and streaming.
- It can be used without LangChain.
Pricing
As written on LangGraph's own page on the read date. Prices change: check there before you pay.
Alternatives to LangGraph
All orchestration tools for agents
Durable execution platform, open source, used to run long-lived agents.

Open source agent framework with SDKs for Python, TypeScript, Go, Java, Kotlin.

Durable execution platform for workflows, background jobs and AI agents.

Typed Python SDK for building agents across many model providers.
Last check: . Method: public repository and docs read automatically, page and day written down. Next check due within 90 days. How we check
Is this your tool? Write to hello@agent-tools.directory from an address on its domain to correct a pin or a fact.