About Agent Framework
Microsoft Agent Framework is the next generation of both Semantic Kernel and AutoGen — combining simple agent abstractions with enterprise-grade capabilities for production AI applications.
What is Microsoft Agent Framework?
Microsoft Agent Framework is an open-source SDK for building AI agent applications in .NET (C#) and Python. It provides a unified, consistent programming model for creating agents that use Large Language Models (LLMs) to reason, plan, and act — with support for function tool calling, multi-turn conversations, streaming responses, and structured output.
The framework is currently in public preview (RC-1) and is available on NuGet
as Microsoft.Agents.AI.OpenAI and related packages.
The History: Semantic Kernel & AutoGen
Two Microsoft projects pioneered the concepts of AI agents in .NET:
- Semantic Kernel — introduced enterprise-grade AI abstractions: kernel plugins, memory, telemetry, and middleware (filters). It focused on integrating LLMs into existing applications with production-quality patterns.
- AutoGen — pioneered simple, composable multi-agent patterns. It made it easy to create agents that collaborate, hand off tasks, and engage in autonomous conversations.
Agent Framework is the direct successor to both, created by the same Microsoft teams. It combines AutoGen's simple agent abstractions with Semantic Kernel's enterprise features, and adds graph-based workflows for explicit multi-agent orchestration.
In short: Agent Framework = AutoGen's simplicity + Semantic Kernel's enterprise features + new graph-based workflows.
Key Capabilities
🤖 Agents
Individual agents that use LLMs to process inputs, call tools and MCP servers, and generate responses. Supports Azure OpenAI, OpenAI, Anthropic, Ollama, and more.
🔀 Workflows
Graph-based workflows that connect agents and functions for multi-step tasks with type-safe routing, checkpointing, and human-in-the-loop support.
🔧 Function Tools
Wrap any C# method as a tool using AIFunctionFactory.Create(). The LLM calls your functions automatically when needed.
🔌 MCP Integration
Built-in support for Model Context Protocol (MCP) — connect to hosted and local MCP tool servers.
⚡ Streaming
Stream agent responses token-by-token with RunStreamingAsync() for real-time, interactive experiences.
📦 Structured Output
Get strongly-typed C# objects from the LLM using RunAsync<T>() — no manual JSON parsing required.
Supported Providers
Agent Framework supports multiple AI providers through a common IChatClient abstraction:
- Azure OpenAI (Chat Completion & Responses API)
- OpenAI (Chat Completion & Responses API)
- Azure AI Foundry (Models + Persistent Agents)
- Anthropic Claude (direct & via Azure Foundry)
- Ollama (local models)
- Any
Microsoft.Extensions.AI.IChatClientimplementation
Current Status
Microsoft Agent Framework is currently in public preview. APIs may change before the final release. Please submit feedback and issues on the GitHub repository.