LangGraph
Open-source library for building agents as explicit graphs of steps and state.
Find alternatives to LangGraph
What it does
- Agent orchestration
Features 5
- Human-in-the-loopAdds moderation and quality controls so people can steer and approve agent actions.
- Customizable agent workflowsLow-level primitives for single, multi-agent and hierarchical control flows in one framework.
- Persistent memoryBuilt-in memory stores conversation histories and maintains context across sessions.
- First-class streamingNative token-by-token streaming showing agent reasoning and actions in real time.
- PersistenceCheckpoints state so agents can resume after failures or pauses.
How it compares
-
CrewAI OSS
CrewAI Flows gives you the same power — event-driven orchestration, conditional routing, shared state — with dramatically less boilerplate and a mental model that maps cleanly to how you actually think about multi-step AI workflows.
Official documentation · 20 Sep 2026 -
Mastra
LangChain's own 'The best AI agent frameworks in 2026' page names Mastra (16 mentions), CrewAI (14) and LangGraph (16) as the frameworks compared.
Company disclosure · 20 Sep 2026 -
deepset Haystack Enterprise Platform
It's most accurate to think of Haystack as covering both LangChain and LangGraph territory.
Official documentation · 22 Sep 2026 -
Elastic Agent Builder
General-purpose frameworks like LangGraph and Amazon Bedrock Agents leave retrieval for you to add and configure.
Official documentation · 13 Oct 2026 -
n8n Cloud
n8n vs. LangGraph: Which is Right for You?
Company disclosure · 13 Oct 2026
Sources
- Pricing
-
The library is open source; its managed deployment is priced on the shared LangChain pricing page.
- Sold within
-
Official documentation · 29 Sep 2026
The vendor's page states a free open-source self-serve route needing no plan; managed deployment is a separate LangSmith product: "LangGraph is an MIT-licensed open-source library and is free to use."
- Interfaces
-
Platform listing · 14 Oct 2026
PyPI listing published by langchain (PyPI maintainer; langchain-ai): 'Building stateful, multi-actor applications with LLMs'.
- Intended for
-
'Developers trust LangGraph to build reliable agents'; customer logos include Expedia, Autodesk, Workday, Nvidia, Coinbase, ServiceNow.
Also from LangChain
-
LangChain Developer tool
Open-source framework for composing language-model calls, tools and retrieval into applications.
-
LangSmith Platform
Traces, evaluates and monitors language-model applications in development and production.
-
LangSmith Context Hub Agent platform
Versions and deploys prompts and agent context. Part of the LangSmith platform (categorized under 'No-Code Agents', 'Observability', 'Evaluation' or 'Deployment infrastructure' depending on the module).
-
LangSmith Deployment Agent platform
Deploys and operates production agents. Part of the LangSmith platform (categorized under 'No-Code Agents', 'Observability', 'Evaluation' or 'Deployment infrastructure' depending on the module).
-
LangSmith Engine Agent platform
Analyzes agent traces and recommends fixes. Part of the LangSmith platform (categorized under 'No-Code Agents', 'Observability', 'Evaluation' or 'Deployment infrastructure' depending on the module); also red-teams agents for issues not yet seen in production.
-
LangSmith Fleet Agent platform
No-code agent building, deployment into Slack/Teams/Gmail, and human-in-the-loop approval controls. Part of the LangSmith platform (categorized under 'No-Code Agents', 'Observability', 'Evaluation' or 'Deployment infrastructure' depending on the module).
-
LangSmith LLM Gateway Agent platform
Governs model calls and spending across providers. Part of the LangSmith platform (categorized under 'No-Code Agents', 'Observability', 'Evaluation' or 'Deployment infrastructure' depending on the module).
-
LangSmith Sandboxes Agent platform
Runs agent-generated code in isolated virtual machines. Part of the LangSmith platform (categorized under 'No-Code Agents', 'Observability', 'Evaluation' or 'Deployment infrastructure' depending on the module).
-
LangSmith Chat Application
LangSmith Chat analyzes application traces, threads, prompts and evaluations; also answers natural-language questions to understand agent performance, debug issues and analyze user sentiment.
-
LangSmith Evaluation Agent platform
LangSmith's evaluation service for agents: run evaluations on curated datasets to compare agent versions and catch regressions, score live traffic with online evaluators, collect human feedback in annotation queues and optimise prompts in a playground.
-
dcode Developer tool
Open-source terminal coding agent built on the Deep Agents SDK that works with the user's own model and lets them control the agent harness, memory, skills, hooks and how code execution is approved, traced and run.
-
Deep Agents Developer tool
Open-source agent harness for long-running tasks that handles planning, context management and multi-agent orchestration, with planning tools, parallel subagents, a virtual filesystem for memory and any model provider.
Open LangChain in the directory
Something wrong here? Send a correction — quote this product id: langgraph.