The official production-ready merger of AutoGen and Semantic Kernel. It provides a unified SDK for multi-agent orchestration, complex workflows, and enterprise-grade observability through Microsoft Foundry.
A library for building stateful, multi-actor applications with LLMs using a graph-based model. It allows for high-control loops, reasoning cycles, and built-in persistent state storage.
Focuses on role-playing autonomous AI agents that collaborate as a "Crew." It excels at multi-agent coordination with simple role-based abstractions and task management.
An open-source, lightweight framework by IBM for building reliable multi-agent systems. It focuses on governance, rule-based constraints, and high parity between Python and TypeScript.
A Python-native agent framework focusing on type safety and structured data. It leverages Pydantic V2 to ensure agent inputs and outputs are strictly validated.