← Back to Discover

babyagi

Experimental Python framework for a self-building autonomous agent: it stores and runs functions from a database, with a dashboard, logging and triggers. Its author says it is not meant for production use.

Autonomous agentsAgent frameworks

What it does

BabyAGI is an experimental framework aimed at a self-building autonomous agent. Its core, functionz, stores, manages and runs functions from a database, tracking imports, dependent functions and secret keys in a graph, with automatic loading and logging. A dashboard lets you register and update functions, view their dependencies, manage secret keys, read execution logs and set up triggers that run one function in response to another. You install it with pip and start the dashboard on your own machine. Two experimental self-building agents are included. One chooses between reusing a function it already has and writing new ones, and the other invents sample tasks for a described user and builds functions to handle them. The README calls these draft features and says the project is not meant for production use. The original March 2023 BabyAGI has been archived in a separate repository.

Capabilities

  • functionz framework for storing and running functions from a database
  • Graph of function dependencies, imports and secret keys
  • Local dashboard for managing functions, keys and logs
  • Logging of every execution with inputs, outputs, timing and errors
  • Triggers that run functions automatically
  • Pre-loaded function packs, including AI-generated descriptions and embeddings

When to use it

  • You want to experiment with self-building agent ideas in Python
  • You want a small, inspectable function registry with a dashboard and execution logs