AutoGen is an open-source programming framework developed by Microsoft Research for building AI agents and facilitating cooperation among multiple agents to solve complex tasks through multi-turn conversations. It pioneered the concept of conversable agents that can interact with each other, use tools, and involve humans in the loop, enabling sophisticated multi-agent workflows for code generation, task planning, and collaborative problem-solving. AutoGen introduced the idea that complex AI tasks are best solved through agent conversations rather than single-shot prompts, fundamentally changing how developers approach agentic AI development.
AutoGen v0.4 represents a complete redesign with a robust, asynchronous, and event-driven architecture that enables broader agentic scenarios with stronger observability, more flexible collaboration patterns, and reusable components. The framework supports various large language models, tool use, autonomous and human-in-the-loop workflows, and multi-agent conversation patterns including group chat, nested conversations, and teachable agents. AutoGen is now part of the broader Microsoft Agent Framework, which merges AutoGen dynamic multi-agent orchestration with Semantic Kernel production foundations for enterprise deployment.
AutoGen targets AI researchers, enterprise developers, and teams building multi-agent applications for software engineering, data analysis, mathematical reasoning, and automated decision-making workflows. It integrates with the Azure AI ecosystem and supports deployment through the Microsoft Agent Framework, which is available in Python and .NET. While AutoGen continues to receive maintenance and security patches, new projects are encouraged to adopt the Microsoft Agent Framework for production use, combining AutoGen orchestration capabilities with Semantic Kernel stability and enterprise tooling.