Lee Stott presents the Microsoft Agent Framework, guiding developers through its unified, open-source SDK for agentic AI apps with .NET and Python support, multi-agent orchestration, and advanced workflow capabilities.

Introducing the Microsoft Agent Framework: Unified SDK for AI Agents and Workflows

The Microsoft Agent Framework is an open-source SDK designed for developers aiming to build intelligent, multi-agent applications in .NET or Python. It delivers a unified foundation by merging the best of Semantic Kernel’s enterprise features with AutoGen’s research-oriented abstractions and introduces advanced capabilities for agent orchestration and workflow design.

Key Features

  • Multi-Agent Orchestration: Simplifies building systems with multiple autonomous agents using both .NET and Python.
  • Integrated Best-of-Breed: Combines AutoGen’s orchestration simplicity and Semantic Kernel’s strong features like thread-based state management, type safety, and telemetry.
  • Graph-Based Workflows: Define complex, modular workflows with routing, conditional logic, checkpointing, and support for human-in-the-loop interventions.
  • Open and Extensible: Developers can extend with middleware, memory context providers, and custom integrations.
  • Supports Major Providers: Out-of-the-box support for Azure OpenAI, OpenAI, and Azure AI.
  • Backward Compatibility: Migration guides ensure a smooth upgrade path from existing Semantic Kernel or AutoGen-based solutions.

Installation and Getting Started

  • Python:

    pip install agent-framework

  • .NET:

    dotnet add package Microsoft.Agents.AI

Integration Ecosystem

  • Works seamlessly with Foundry SDK, MCP SDK, A2A SDK, and integrates with M365 Copilot Agents.
  • Rich library of declarative agent manifests and code samples.
  • Learning resources and community involvement provided through Microsoft Learn modules, GitHub repositories, and the Azure AI Foundry Discord server.

Workflow and Agent Design

  • AI Agents: Enable dynamic, decision-making components that process inputs, leverage state management, call tools or MCP servers, and generate responses.
  • Workflows: Structure complex processing using type-based routing, orchestration patterns (sequential, concurrent, hand-off), and checkpoints for reliability and flexibility.

Example Use Cases

  • Building enterprise customer support bots
  • Automating research agent tasks
  • Orchestrating code generation or educational AI systems
  • Integrating human oversight into process loops

Migration and Compatibility

Developers using Semantic Kernel or AutoGen will find migration guides and backward compatibility built in. Community input is welcome with active feedback and contributions on GitHub.

Important Notes

  • Agent Framework is currently in public preview; developers are encouraged to participate in testing and to be mindful of data handling and compliance when integrating third-party agents or services.
  • Documentation, downloads, and migration resources available via official Microsoft resources.

Community and Support Resources

Get started by downloading the SDK, exploring documentation, and joining the growing developer community building the next generation of AI agentic workflows.

This post appeared first on “Microsoft Tech Community”. Read the entire article here