AutoGen

AutoGen (from Microsoft and the community) models applications as conversations among agents—LLM-backed assistants, human participants, and tool executors—that coordinate to solve coding, analysis, or planning tasks. It supports flexible speaker selection, code execution sandboxes, and integration with popular model APIs.

The project is open source and actively developed, with documentation oriented toward researchers and engineers experimenting with agentic patterns at scale. It complements rather than replaces raw SDK calls by supplying conversation graphs and orchestration primitives.

Teams choose AutoGen when they want explicit multi-party dialogue semantics and Microsoft-backed examples for software engineering and data workflows.