Tools · SPBoucher
Menu

LangGraph

In-depth entry

by LangChain, Inc. · Framework · also known as langgraph (PyPI), LangGraph.js, LangGraph Platform (former deployment name)

Low-level orchestration framework and runtime for stateful, long-running agents expressed as graphs, with durable execution, checkpointing, human-in-the-loop and streaming

MIT licenseOpen sourceFreeFreemiumEnterpriseLocalSelf-hostedCloudSelf-hostable
Visit official site Compare

Link verified Nov 2, 2026

Overview

LangGraph is an MIT-licensed framework and runtime from LangChain, Inc. for building long-running, stateful agents. Rather than hiding the agent loop behind a prompt abstraction, it asks the developer to define a StateGraph: typed state, nodes that are ordinary functions or LLM calls, and edges (fixed or conditional) that connect them to START and END before compile() produces a runnable. This lets deterministic steps and model-driven steps live in one graph, and it can be used with or without the LangChain library. Python and JavaScript implementations are maintained in parallel.

The runtime is where most of the value sits. Checkpointers persist state after every step, so a run can survive a crash and resume where it stopped; the same mechanism provides time travel for debugging, short-term memory within a thread and long-term memory stores across threads. Interrupts pause execution so a human can inspect or edit state before approving a tool call, and streaming surfaces tokens, state updates and events as they happen. LangSmith adds tracing and a visual studio, and the hosted path, now called LangSmith Deployment (previously LangGraph Platform), runs graphs as managed services with cron scheduling, MCP exposure and an Assistants API; self-hosted and hybrid variants exist on the Enterprise plan.

LangGraph suits workflows where reliability matters more than convenience: multi-step research agents, approval-gated automations and anything that must pause for hours. The cost is verbosity and a learning curve around state schemas and reducers, and production-grade persistence requires a database-backed checkpointer rather than the in-memory default.

Capabilities

  • Graph definition with typed state, nodes, conditional edges and subgraphs
  • Durable execution with checkpointers and resumable runs
  • Human-in-the-loop interrupts to review or edit state
  • Short- and long-term memory stores
  • Token, update and event streaming
  • Python and JavaScript libraries usable without LangChain
  • LangSmith tracing and studio; hosted LangSmith Deployment with cron, MCP and Assistants API
  • Prebuilt agent helpers and the Deep Agents SDK built on top

Typical workflows

  1. Model an approval-gated automation that pauses for a human before side effects
  2. Build a multi-step research agent with persistent memory across sessions
  3. Deploy a graph as a managed service with scheduled runs
  4. Debug a failed run by replaying from a checkpoint

Strengths

  • Explicit control flow instead of hidden prompts
  • Persistence and interrupts are first-class
  • Framework-agnostic model and tool usage
  • MIT license with optional commercial hosting

Limitations

  • More code than high-level agent frameworks
  • In-memory checkpointer is not production-grade; a database backend is needed
  • Deployment product has been renamed, which dates older tutorials
  • Hosted features beyond the free tier require LangSmith plans

Prerequisites

  • Python 3.10+ or Node.js
  • A model provider or local runtime; a Postgres or SQLite checkpointer for durable runs

Who it is for

  • Engineers building production agents with reliability requirements
  • Teams needing human approval steps in automations
  • Researchers prototyping multi-agent workflows

Alternatives & comparison

LangChain's create_agent is a higher-level harness that runs on LangGraph; use LangGraph directly when you need custom control flow. Pydantic AI emphasizes typed outputs and dependency injection and offers durable execution through Temporal or DBOS rather than its own checkpointer. The OpenAI Agents SDK is lighter and tied to OpenAI's Responses API by default. CrewAI models work as role-based crews and flows. AutoGen is archived and points users to Microsoft Agent Framework.

  • AutoGen

    Microsoft's multi-agent framework (Core, AgentChat, Extensions, Studio) for Python and .NET, now archived with users directed to Microsoft Agent Framework

    Open sourceOpen source
  • CrewAI

    Open-source Python framework for role-based multi-agent crews and event-driven flows, with a commercial enterprise console for deploying, monitoring and triggering automations

    MixedOpen source
  • LangChain

    Open-source Python and JavaScript framework whose create_agent harness builds tool-using agents on the LangGraph runtime with swappable model providers and middleware

    Open sourceOpen source
  • LlamaIndex

    Open-source Python framework for retrieval-augmented agents over your data, with connectors, indexes, query engines and event-driven workflows, alongside the commercial LlamaParse document platform

    Open sourceOpen source
  • OpenAI Agents SDK

    Lightweight open-source Python and TypeScript framework for agents, handoffs, guardrails, sessions and tracing, defaulting to OpenAI's Responses API with adapters for other providers

    Open sourceOpen source
  • Pydantic AI

    Type-safe Python agent framework from the Pydantic team with validated structured outputs, dependency injection, swappable model providers, MCP, durable execution and evals

    Open sourceOpen source
Compare LangGraph with its alternatives →

Sources & verification

Primary sources consulted for this record. Claims are tied to the fields each source supports; the automated link checker re-verifies every URL on a schedule.

Confidence
High
Reviewed
Nov 1, 2026
Links
Verified · Nov 2, 2026

Spotted an error?

Broken link, outdated license, wrong category, or a tool that no longer exists? Tell us; reports are reviewed by the editor.

Apify

Apify Technologies s.r.o. · Data collection & crawling

In-depth entry

Cloud platform and marketplace for running web-scraping and automation programs (Actors), with storages, proxies, scheduling, an API and an MCP server for agents

Apache-2.0FreemiumUsage-basedNov 2, 2026

Claude Code

Anthropic · Coding agents & AI IDEs

In-depth entry

Anthropic's agentic coding tool that reads a codebase, edits files and runs commands from the terminal, IDEs, a desktop app, the web and CI, with MCP, hooks, skills and subagents

ProprietaryPaidUsage-basedLocalNov 2, 2026

Claude Platform

Anthropic · Model providers & APIs

In-depth entry

Anthropic's developer platform for the Claude models: Messages API, tool use, structured outputs, prompt caching, batch processing and Managed Agents, plus cloud availability

ProprietaryUsage-basedEnterpriseNov 2, 2026

OpenAI API

OpenAI · Model providers & APIs

In-depth entry

OpenAI's hosted developer platform: Responses and Chat Completions APIs, Realtime voice, embeddings, images, audio, fine-tuning, evals and managed agents

ProprietaryUsage-basedEnterpriseNov 2, 2026

Model Context Protocol

Model Context Protocol, a Series of LF Projects, LLC (originated by Anthropic) · Agent frameworks & orchestration

In-depth entry

Open JSON-RPC standard for connecting AI applications to tools, data and prompts through servers, now stewarded under the Linux Foundation with SDKs in many languages

Open sourceOpen sourceFreeLocalSelf-hostedNov 2, 2026

LlamaIndex

LlamaIndex, Inc. · Agent frameworks & orchestration

Open-source Python framework for retrieval-augmented agents over your data, with connectors, indexes, query engines and event-driven workflows, alongside the commercial LlamaParse document platform

MITOpen sourceFreeLocalSelf-hostedNov 2, 2026