← Back to Discover

AutoGen

Python framework for building applications where several AI agents cooperate, with AgentChat and Core APIs and a no-code Studio. Archived by Microsoft and no longer maintained; Microsoft Agent Framework is the named successor.

Agent frameworksAutonomous agents
Visit website ↗

This product has been discontinued.

What it does

AutoGen is a Python framework for building applications where several AI agents cooperate, either independently or with a person in the loop. Its design is layered: a Core API for message passing and event-driven agents, an AgentChat API for rapid prototyping of patterns such as two-agent and group chats, and an Extensions API for model clients and code execution. Related tools include AutoGen Studio, a no-code interface for prototyping multi-agent workflows, and AutoGen Bench, a benchmarking suite for agents. The README shows agents using OpenAI models and an MCP server for web browsing, and AutoGen requires Python 3.10 or later. Microsoft states that the project is archived and no longer maintained, with no bug fixes, security fixes, new features or support, and it advises against production use. New users are pointed to Microsoft Agent Framework, and a migration guide is provided for existing users.

Capabilities

  • Layered Core, AgentChat and Extensions APIs
  • Multi-agent patterns including two-agent chat and group chats
  • AutoGen Studio, a no-code interface for prototyping workflows
  • MCP tool support and model clients such as OpenAI's
  • AutoGen Bench for benchmarking agent performance

When to use it

  • You maintain an existing AutoGen project and need reference material or a migration path
  • You want to study an early multi-agent orchestration design for research or experiments

Alternatives to explore

  • Microsoft Agent Framework