AutoGen is an open-source framework from Microsoft that makes it easy to build complex workflows using large language models like GPT-4. It allows you to create customizable “agents” – AI assistants, humans, tools, or a mix – that can have conversations with each other to solve tasks.

AutoGen Key Features

  • Lets you quickly build agents with different capabilities using GPT-4, humans, and tools. Agents can take on specialized roles and have back-and-forth conversations.
  • Supports features like automated inference tuning, integrating human input, and invoking functions. Enables new creative applications like conversational chess.
  • Reduces the effort needed to orchestrate optimal workflows with LLMs. Cuts down coding time significantly compared to doing this from scratch.
  • Available as an easy-to-use Python package. Enables powerful experiences in just a few lines of code.