About AutoGen
AutoGen is an open source framework from Microsoft for building applications that use one or more AI agents. It supports simple single-agent setups as well as larger multi-agent systems.
What it does
AutoGen is organized in layers for different needs. AgentChat is used for conversational agents in Python, Core handles event-driven systems, and Extensions connect to outside services such as MCP servers and Docker-based code execution. A web-based studio lets you prototype agents without writing code.
Key features
- Multi-agent conversations
- Event-driven core layer
- MCP server support
- Code execution in Docker containers
- No-code studio for prototypes
Who it's for
Python developers and research teams who want to build or test multi-agent workflows. Teams new to agents may start with the studio before writing any code.
Pricing and good to know
AutoGen is open source and installs with pip. Model costs come from whichever provider you connect. Start with AgentChat for most conversational projects, and move to Core only when you need more control over message flow.
Highlights
- Multi-agent conversations
- Event-driven core layer
- MCP server support
- Code execution in Docker containers
- No-code studio for prototypes
