AgenticFrameworks

LangChain Academy: Introduction to LangGraph

by LangChain

IntermediateCourseFree~6 hours of video across 55 lessons, self-paced

Six hours and 55 lessons from the LangGraph team — state, memory, breakpoints and deployment.

Start LearningAdded Jul 4, 2026 · Updated Aug 6, 2026

Overview

LangChain Academy's Introduction to LangGraph is the vendor's own free training for its agent-orchestration framework, listed on the platform as "Foundation: Introduction to LangGraph - Python" and running about six hours of video across 55 lessons. It is structured in seven modules. Module 0 is environment setup; Module 1 builds from a simple graph through LangSmith Studio, chains, routers and a first agent with memory; Module 2 covers state schemas, state reducers, multiple schemas, trimming and filtering messages, and a chatbot that summarizes its own history into internal and external memory; Module 3 covers streaming, static and dynamic breakpoints, editing state for human feedback, and time travel; Module 4 builds a research assistant using parallelization, sub-graphs and map-reduce; Module 5 covers short- versus long-term memory, the LangGraph Store, and memory schemas for profiles and collections; Module 6 covers deployment concepts, creating and connecting to a deployment, double texting and assistants. The companion repository, langchain-ai/langchain-academy, is MIT-licensed with roughly 2,800 stars, was last updated in June 2026, and pairs each lesson with a Jupyter notebook plus a studio subdirectory you run locally with langgraph dev. Budget for the real cost of taking it: Python 3.11, 3.12 or 3.13, an OpenAI API key with credits, a LangSmith account for tracing, and optionally a Tavily key for Module 4. It is vendor training, so it teaches the LangChain way rather than comparing frameworks.

At a Glance

Topic
Agentic
Level
Intermediate
Format
Course
Cost
Free
Duration
~6 hours of video across 55 lessons, self-paced
Provider
LangChain
Hands-on
Yes — code/exercises
Certificate
None

What You’ll Learn

  • Model an agent as a graph of nodes, edges and conditional routing
  • Design state schemas and reducers so parallel branches merge correctly
  • Trim, filter and summarize message history to control context growth
  • Add streaming, static and dynamic breakpoints for human-in-the-loop review
  • Edit graph state mid-run and time-travel back to an earlier checkpoint
  • Build long-term memory with the LangGraph Store, profiles and collections
  • Deploy a graph as a hosted assistant and handle double-texting

Highlights

  • Six hours and 55 lessons — substantially deeper than a typical vendor short course
  • Written and maintained by the LangGraph team, so it tracks the current API
  • Every lesson ships as a notebook plus a local LangGraph Studio graph you can run
  • The companion repository is MIT-licensed and was last updated in June 2026
  • Honest caveat: needs a funded OpenAI key and a LangSmith account, and teaches only the LangChain way

Who It’s For

Best For

  • Teams standardizing on the LangChain and LangGraph stack
  • Engineers who need human-in-the-loop control rather than an autonomous agent loop
  • Developers who have outgrown a single-prompt chain and need explicit state

Prerequisites

  • Intermediate Python, plus comfort with virtual environments and Jupyter
  • Python 3.11, 3.12 or 3.13 installed locally
  • An OpenAI API key with credits and a LangSmith account for tracing

FAQ

What is LangChain Academy: Introduction to LangGraph?

LangChain's own free six-hour course on LangGraph, its framework for stateful, controllable agents. Fifty-five lessons across seven modules take you from a two-node graph to a deployed research assistant with long-term memory, human-in-the-loop breakpoints and time travel, each paired with a runnable notebook. It is maintained by the team that ships LangGraph, so it tracks the current API closely.

Is LangChain Academy: Introduction to LangGraph free?

LangChain Academy: Introduction to LangGraph is free to access.

What level is LangChain Academy: Introduction to LangGraph for?

LangChain Academy: Introduction to LangGraph is aimed at a intermediate audience. Recommended background: Intermediate Python, plus comfort with virtual environments and Jupyter, Python 3.11, 3.12 or 3.13 installed locally, An OpenAI API key with credits and a LangSmith account for tracing.

How long does LangChain Academy: Introduction to LangGraph take?

Expect roughly ~6 hours of video across 55 lessons, self-paced. Most learners work through it at their own pace.

What will I learn from LangChain Academy: Introduction to LangGraph?

You'll learn: Model an agent as a graph of nodes, edges and conditional routing; Design state schemas and reducers so parallel branches merge correctly; Trim, filter and summarize message history to control context growth; Add streaming, static and dynamic breakpoints for human-in-the-loop review; Edit graph state mid-run and time-travel back to an earlier checkpoint; Build long-term memory with the LangGraph Store, profiles and collections; Deploy a graph as a hosted assistant and handle double-texting.

Topics

LangGraphLangChainagentsmemoryhuman-in-the-loop

Sources

This page was written from 2 sources, 1 on domains other than academy.langchain.com.

  1. 1.academy.langchain.comintro to langgraphvendor
  2. 2.github.comlangchain academy