BabyAGI is an open-source AI-powered task management system that uses OpenAI and vector databases to create, prioritize, and execute tasks autonomously. It operates on a continuous loop, pulling tasks from a list, sending them to an execution agent powered by OpenAI’s API, and then enriching and storing the results in vector databases like Chroma or Weaviate for context. New tasks are created based on the outcomes of previous ones and a predefined objective. The system is designed to be simple, allowing users to understand and build upon it easily. It supports all OpenAI models, including Llama, and is designed to run continuously as part of a task management system, although users are cautioned about potential high API usage. BabyAGI represents a simplified version of a Task-Driven Autonomous Agent, distilled down to a concise script, and is part of a broader community effort to explore various approaches to expanding autonomous AI systems.
https://github.com/yoheinakajima/babyagi