AutoGen
AutoGen is an open-source framework from Microsoft Research for building applications powered by multiple LLM agents that can converse with each other, call tools, and involve humans in the loop to complete complex tasks.
Definition
AutoGen is an open-source framework from Microsoft Research for building applications powered by multiple LLM agents that can converse with each other, call tools, and involve humans in the loop to complete complex tasks.
Overview
Instead of a single prompt-and-response exchange, AutoGen lets developers define multiple agents with distinct roles — for example, a planner, a coder, and a critic — that exchange messages to collaboratively work through a task, optionally executing code or calling external tools along the way. It is built to work with models from providers such as OpenAI and Azure OpenAI, and is commonly used to automate multi-step coding, research, and data-analysis tasks where a single model call falls short. It sits alongside frameworks like LangChain in the broader agentic AI landscape covered in the AI Agents & Agentic Workflows course and the blog post AI Agents Explained: How They Actually Work. AutoGen was released by Microsoft Research in 2023 as open source, part of Microsoft's broader push into agentic AI tooling.
Key Features
- Multi-agent conversation framework where agents exchange messages to solve tasks
- Built-in code-execution agent for running and testing generated code
- Human-in-the-loop support for approving or steering agent actions
- Configurable agent roles such as planner, executor, and critic
- Works with OpenAI, Azure OpenAI, and other LLM backends
- Open source and extensible with custom tools and agent behaviors