BabyAGI

The minimal task-driven autonomous agent that helped define the genre.

BabyAGI is an experimental, intentionally tiny autonomous agent that uses an LLM plus a vector store to spawn, prioritize, and execute a list of tasks toward an objective. Released alongside AutoGPT in the early autonomous-agent wave, its few-hundred-line core became a canonical reference implementation of the task-loop pattern. It has since evolved into a framework for experimenting with self-building autonomous agents and function packs.