microsoft/agent-framework

What it solves

Microsoft Agent Framework (MAF) provides a consistent, production-grade foundation for building and operating AI agents and multi-agent workflows. It bridges the gap between simple prototypes and production systems by offering tools for orchestration, durability, and observability that are often missing in basic chat loops.

How it works

MAF is a multi-language framework supporting both Python and .NET. It allows developers to define agents and orchestrate them using graph-based patterns—such as sequential, concurrent, and group collaboration—to handle complex tasks. The framework supports multiple LLM providers (including Azure OpenAI and OpenAI) and includes a middleware system for custom request/response processing. It also offers declarative agent definitions via YAML and a dedicated DevUI for testing and debugging.

Who it’s for

It is designed for development teams who need to move beyond stateless chat loops to build durable, observable, and governed agentic systems that can be deployed to local or cloud environments.

Highlights

• Multi-language support: Consistent APIs across Python and C#/.NET.
• Graph-based orchestration: Supports complex workflows including handoffs, concurrent execution, and human-in-the-loop control.
• Production readiness: Built-in OpenTelemetry integration for monitoring and features like checkpointing and time-travel.
• Flexible deployment: Supports various hosting patterns, including Foundry-hosted agents, Azure Functions, and Durable Task hosting.
• Extensible architecture: Includes a middleware system and support for multiple LLM providers to prevent vendor lock-in.