Microsoft AutoGen
Framework for building multi-agent AI systems where LLM agents collaborate, debate, and use tools to solve complex tasks — from Microsoft Research
Visit Microsoft AutoGen →AutoGen structures multi-agent systems as a set of conversable agents that message each other — a task gets broken down and passed between agents (a planner, a coder, a critic reviewing output) that debate and refine a solution rather than one model handling everything in a single pass.
The conversational framing is what distinguishes it from graph-based frameworks like LangGraph — AutoGen agents talk to each other in a chat-like loop, rather than following an explicit, pre-defined workflow graph.