Microsoft Agent Framework Reaches Stable Release After Years of Development

Microsoft has finally settled its Agent Framework into a stable release after years of development. The framework, which combines the features of Semantic Kernel and AutoGen, is now available in version 1.19.0 for Python and 1.22.0 for .NET. According to Microsoft, the framework provides an open-source SDK for building AI agents and multi-agent workflows.

The Agent Framework allows developers to create agents that can call tools and generate responses using LLM-backed workers. It also supports explicit graph-based workflows for controlling how agents and functions execute in sequence. The framework has four main components: Agents, Harness Agents, Workflows, and Integrations.

Microsoft describes the Agent Framework as a direct successor to both Semantic Kernel and AutoGen, built by the same internal teams. The framework is MIT-licensed and hosted on GitHub, where it has accumulated over 13,800 stars. It supports various model providers, including OpenAI, Azure OpenAI, Microsoft Foundry, Anthropic, Amazon Bedrock, Google Gemini, and Ollama.