Microsoft has released version 1.0 of the Agent Framework, its production-ready open-source SDK for building, orchestrating, and deploying AI agents. The release, announced on April 3, fulfills a goal set when the project launched last October: merge the enterprise foundations of Semantic Kernel with the multi-agent orchestration patterns of AutoGen into a single SDK.

What's New in 1.0

The 1.0 release ships with stable APIs and a long-term support commitment across both Python and .NET. Key capabilities include:

  • Graph-based workflows — connect agents and deterministic functions with streaming, checkpointing, and human-in-the-loop support
  • Multi-provider support — first-party connectors for Microsoft Foundry, Azure OpenAI, OpenAI, Anthropic Claude, Amazon Bedrock, Google Gemini, and Ollama
  • A2A + MCP interoperability — agents can communicate across runtimes using the Agent-to-Agent and Model Context Protocol standards
  • DevUI — an interactive developer interface for building, testing, and debugging workflows visually

Getting started takes under ten lines of code in either language. A sequential multi-agent workflow — where one agent drafts content and another reviews it — requires roughly 30 lines.

Why It Matters

The framework consolidates what was a fragmented ecosystem. Developers previously had to choose between Semantic Kernel's enterprise tooling and AutoGen's experimental multi-agent patterns. Agent Framework 1.0 makes both available under a unified abstraction with backward compatibility guarantees.

Migration guides from both Semantic Kernel and AutoGen are included in the documentation.

The project is available on PyPI and NuGet.