Plausible Analytics“We picked LangGraph over CrewAI and AutoGen because it gives us explicit, stateful agent graphs with checkpoints and human approval steps, so we control every step instead of hoping the agents behave.”
AI SDK & Agent FrameworkMaker says so · source ↗
LangGraph
Open-source framework from the LangChain team for building stateful, long-running agents as graphs of steps, with persistence, human-in-the-loop and streaming, in Python and JavaScript.
Works with AI agents:llms.txtCLIAsk your AI about this, with this page as the source:ChatGPT ↗Claude ↗Perplexity ↗
LangGraph is an open-source orchestration library for agents that run many steps, run for a long time, or need a person to step in. You describe the agent as a graph: nodes are functions (an LLM call, a tool call, or ordinary code), edges decide what runs next, and a shared state object flows between them. Deterministic steps and model-driven steps live in the same graph.
The model a developer has to understand is the state graph plus checkpoints. You define a state schema, add nodes and edges, and compile the graph. Compiled with a checkpointer, LangGraph saves the state after each step under a thread id, which is what lets a run pause for human approval, resume after a crash, or be rewound. Libraries exist for Python and JavaScript/TypeScript; LangChain is optional.
What a small team gets: durable execution, streaming of intermediate steps, interrupts for human-in-the-loop, short-term memory per thread and a separate store for long-term memory across threads. Checkpointers come for SQLite (local work) and Postgres or MongoDB (production). It works with any model, including open-weight ones. You run it in your own app, or deploy it on LangSmith Deployment, which runs as cloud, bring-your-own-cloud or self-hosted.
It is deliberately low-level: it does not hand you prompts or an agent architecture. The LangChain team points people who want a prebuilt tool-calling loop to LangChain's agents instead. Without LangSmith, you run the persistence database, scaling and retries yourself.
Where it fits
How LangGraph itself is built
1 tools, from its own code, website and Product Hunt page.
Who uses it
1 makers' products, each linked to the source that shows it, and 87 open-source projects that declare it in their code.
NovuNotification infrastructure for sending multi-channel messages (email, in-app, SMS, chat…AI SDK & Agent FrameworkIn its code · source ↗
omiOpen-source AI wearable with mobile and desktop apps that records conversations and meet…AI SDK & Agent FrameworkIn its code · source ↗
PostHogProduct analytics with session replay, feature flags, A/B tests and surveys in one open-…AI SDK & Agent FrameworkIn its code · source ↗
PostizSocial media scheduling and analytics tool for planning, AI-assisted drafting, and auto-…AI SDK & Agent FrameworkIn its code · source ↗
DeepEvalOpen-source Python framework for unit-testing LLM apps with pytest-style test cases and…AI SDK & Agent FrameworkIn its code · source ↗Open source: a project that declares LangGraph as a dependency in its public code — verifiable, but not necessarily a live product.
What makers say
1 makers on why they use LangGraph, in their own words on Product Hunt.
We picked LangGraph over CrewAI and AutoGen because it gives us explicit, stateful agent graphs with checkpoints and human approval steps, so we control every step instead of hoping the agents behave.
Loved and watch-outs
Themes that recur in makers' words and Hacker News comments, each linked to what it summarises.
Most loved
- Explicit stateful graphs with checkpointing and human-in-the-loop steps give control over every step of an agent. PHHNHN 2
- Nodes put fences around what the model may do, so a defined workflow is followed reliably. HNHN 2
- Checkpointing and state management save teams from the edge-case bugs of a homegrown version. HNHN 2
Reliability and open issues
Most wanted on GitHub
- langgraph-checkpoint-postgres (psycopg.OperationalError: sending query and params failed: SSL error: bad length) encountered across multiple version👍 12 · opened Mar 2025 · active Aug 2026
Who switches
Public pull requests on GitHub since Oct 2024 whose title says "X to Y" — real code changes moving a project from one tool to another, by developers in general. Open a row to see the pull requests.
LangChain → LangGraph6+ PRs
- feat(langchain): bridge MCP elicitation to LangGraph interruptslangchain-ai/langchain · 2026-08-20
- Migrate LLM calls to LangChain + orchestration to LangGraphKomali612/cicd-bootstrap · 2026-08-07
- archiving current langchain demo to start langgraph implementationyajji-k/ai-engineering-workspace · 2026-07-30
- Shifted chatbot form langchain to langgraph. Improved the workflow, M…Nouman64-cat/insurance-ai · 2026-07-23
- docs: add missing langchain to the LangGraph Python install commandComposioHQ/composio · 2026-07-15
- feat: migrate routing workflow from langchain to langgraphChikkalaSukrutiNaidu/multimodal-rag-bot · 2026-06-15
LangGraph → Mastra5 PRs
- feat(mastra): migrate v2 LangGraph chat extension to Mastra v3sourcefuse/loopback4-llm-chat-extension · 2026-05-22
- feat(migration): migrate lb4-llm-chat-extension from LangGraph to Mastra sourcefuse/loopback4-llm-chat-extension · 2026-05-13
- Migrate agent from LangGraph to MastraQuenumGerald/NudgeBot · 2026-05-07
- feat(migration): migrate lb4-llm-chat-extension from LangGraph to Mastra [WIP]sourcefuse/loopback4-llm-chat-extension · 2026-05-04
- Plan langgraph to mastra migrationConnorbelez/open-canvas · 2025-09-01
LangGraph → LangChain3+ PRs
- Slim down qvac-langgraph to a pure LangChain adapterSpaceUY/qvac-meridian-challenge · 2026-10-01
- docs(cli): rename langgraph template to langchain in package READMEsassistant-ui/assistant-ui · 2026-07-04
- feat(cli): rename the langgraph starter template to -t langchainassistant-ui/assistant-ui · 2026-06-19
Alternatives to LangGraph
All alternatives by situation →
LlamaIndex
OpenAI Agents SDK
Pydantic AI
AI SDK
Claude Agent SDK
CrewAIQuestions makers ask about LangGraph
Do I need LangChain to use LangGraph?
No. LangGraph is a lower-level orchestration library and can be used on its own; LangChain components are optional helpers for models and tools. source ↗
Does it work with open-source models or models without tool calling?
Yes. It does not depend on a particular model. With a model that lacks tool calling you write a little code to turn its text output into the next decision. source ↗
What do I need to self-manage if I don't use LangSmith?
The open-source library leaves persistence, deployment, scaling, retries and monitoring to you, and has no HTTP API or cron scheduling. LangSmith Deployment adds managed Postgres persistence, auto-scaling task queues, cron and an HTTP API for state. source ↗
Which databases can store agent state?
Official checkpointers cover in-memory, SQLite (recommended for experiments), and Postgres or MongoDB for production; an Azure Cosmos DB checkpointer is also available. Each is a separate package. source ↗
Can I deploy it on my own infrastructure?
Yes. You can run the library inside any app you host, or use LangSmith Deployment on cloud, bring-your-own-cloud or your own infrastructure, all using the same Agent Server runtime. source ↗
Can I use the visual Studio debugger without a LangSmith account?
Yes. Run the local development Agent Server and connect it to Studio; set LANGSMITH_TRACING=false and no traces are sent to LangSmith. source ↗
Is JavaScript a first-class option?
Yes. LangGraph has Python and JavaScript/TypeScript libraries with the same core concepts; the JS package is @langchain/langgraph. source ↗
Is LangGraph free?
Yes — there is a free tier a small product can run on; paid use starts at $39/seat/mo (LangSmith Plus, includes one small deployment). source ↗
Is LangGraph open source or self-hostable?
Open source, and you can self-host it. source ↗
Can AI coding agents work with LangGraph?
It serves an llms.txt docs index; it has an official CLI (@langchain/langgraph-cli).
Who uses LangGraph?
1 makers' products we track, each with a source, and 87 open-source projects declare it in their code. source ↗