Microsoft Agent Framework is 1.0! Multi-agent workflows and orchestration, A2A + MCP, works with any model provider - across .NET and Python. From zero to agent in 5 lines of code. aka.ms/AgentFramework

DevBlogs is a curated collection of articles and blog posts from developers and tech enthusiasts worldwide, covering a wide range of topics such as programming, software engineering, and technology trends. Readers can learn about new programming techniques, software architecture patterns, and industry trends from diverse perspectives, helping them stay informed and inspired in their software development journey.

DevBlogs

Microsoft Agent Framework has reached version 1.0 for both .NET and Python, marking a production-ready release with stable APIs and long-term support commitment. The framework unifies Semantic Kernel and AutoGen into a single open-source SDK. Key features include single and multi-agent orchestration patterns (sequential, concurrent, handoff, group chat, Magentic-One), pluggable memory backends (Redis, Neo4j, Mem0), middleware hooks, declarative YAML-based agent definitions, MCP tool integration, and A2A cross-runtime agent collaboration. First-party connectors support Azure OpenAI, OpenAI, Anthropic Claude, Amazon Bedrock, Google Gemini, and Ollama. Preview features include a browser-based DevUI debugger, GitHub Copilot SDK and Claude Code SDK integration, AG-UI/CopilotKit adapters, and reusable Skills packages. Migration assistants are available for teams coming from Semantic Kernel or AutoGen.

Microsoft Agent Framework Version 1.0