Microsoft has introduced the Microsoft Agent Framework, an open source SDK and runtime for building, tuning and deploying AI agents and multi-agent workflows with full framework support for .NET and Python.
The agent framework, introduced on October 1st and available on GitHub, is designed to allow developers to build everything from simple chat agents to complex multi-agent workflows using graph-based orchestration. Developers experiment with the framework locally and deploy it to AI AI Foundry for AI apps, with built-in observability, durability and compliance. Multi-agent systems can connect Azure AI Foundry, Microsoft 365 Copilot, and other agent platforms.
[ Related: Agentic AI – Ongoing news and insights ]
Microsoft said the Microsoft Agent Framework has extended ideas from the semantic kernel and Autogen project, and developers currently using these projects will find a simple move to the Agent Framework. The agent framework supports the Model Context Protocol (MCP) and Agent2Agent (A2A) protocols, and features Openapi-First Design and Cloud-Ingostic Runtime. As a result, Microsoft says the agent framework enables the following features:
