Agentic Learning StudioGenerate your own lesson →
AI Agent Frameworks: LangChain, LangGraph, CrewAI and More

Frameworks · Guide

AI Agent Frameworks: LangChain, LangGraph, CrewAI and More

Pick the right scaffolding — LangChain, LangGraph, CrewAI, AutoGen, LlamaIndex, OpenAI Agents SDK.

Agent frameworks give you the scaffolding — tool calling, state management, and orchestration — so you don't rebuild the agent loop by hand. This guide compares LangChain, LangGraph, CrewAI, AutoGen, LlamaIndex, the OpenAI Agents SDK, and Claude tool use, and helps you choose the right one for your project.

Generate your own lesson →

What you'll learn

Lessons in this guide (11)

Frequently asked questions

LangChain vs LangGraph — what's the difference?

LangChain is a broad toolkit of components (models, prompts, tools, chains). LangGraph is its framework for building agents as an explicit, inspectable state graph you fully control — better when you need branching, loops, and durable state.

Which agent framework should I use?

For a simple tool-using loop, start with plain tool calling or LangChain. For complex, stateful, multi-step agents, use LangGraph. For role-based multi-agent teams, CrewAI or AutoGen. Match the framework to how much control and structure you need.

What is CrewAI?

CrewAI is a framework for orchestrating multiple role-based agents (e.g. researcher, writer, reviewer) that collaborate on a task, run sequentially or hierarchically.

Other guides