Microsoft updated his AutoGen orchestra framework so the agents it helps build can be more flexible and give organizations more control.

AutoGen v0.4 brings robustness to AI agents and addresses issues identified by customers around architectural constraints.

“The initial release of AutoGen generated widespread interest in agent technology,” said Microsoft researchers. in a blog post. “At the same time, users struggle with architectural constraints, a poor API combined with rapid growth and limited debugging and intervention functionality.”

The researchers added that customers are asking for stronger monitoring and control, flexibility around multi-agent collaboration and reusable components.

AutoGen v0.4 is more modular and extensible, with scalability and distributed agent networks. It adds asynchronous messaging; cross-language support, observation and debugging; and built-in and community extensions.

Asynchronous messaging means agents built with AutoGen v0.4 support event-driven and request-interaction patterns. The framework is more modular, so developers can add plug-in components and build long-running agents. It also allows users to design more complex and distributed agent networks.

The AutoGen extension module simplifies the process of working with multi-agent teams and advanced model clients. It also allows open-source developers to manage their extensions.

To address the observability issue, AutoGen v0.4 has built-in metric tracking, messaging tracing and debugging tools so users can monitor agent interactions. The updates enable interoperability between agents that speak different coding languages; currently, AutoGen v0.4 supports Python and .NET, but support for additional languages ​​is in the works.

New framework

Microsoft has updated the AutoGen framework to better define responsibilities across the framework, tools and applications.

It has three layers: core, which contains the foundational building blocks for an event-driven system; AgentChat, a “task-driven, high-level API built in the core layer” with group chat, code execution and pre-built agents and most similar to AutoGen v0.2; and first-party extensions, which interface with integrations such as the Azure code executor and the OpenAI client model.

Along with its framework update, some of the tools Microsoft built around AutoGen also got an upgrade.

AutoGen Studio, a low-code interface for rapid prototyping agents, has been rebuilt with the AutoGen v4.0 AgentChat API. Users can get real-time agent updates, pause conversations or redirect agents with mid-execution control, agent design teams with a drag-and-drop interface -and-drop, import custom agents and get interactive feedback.

Microsoft and agents

Released by Microsoft AutoGen in October 2023 with the hope of simplifying how agents communicate with each other. Along with LangChain and LlamaIndexAutoGen was one of the first AI agent orchestration frameworks released before agents became the buzzword they are today.

Since then, Microsoft has released other agent systems including Magentic-Onea generalist agentic system that can empower multiple agents to complete tasks.

The company accepts AI agents, deploying perhaps the largest AI agent ecosystem through the Copilot Studio platform.

But other companies are hot on their heels. Salesforce AgentForce launched, and its new updated AgentForce 2.0while ServiceNow released a library of customizable agents. AWS also added more support for creating multi-agent systems in its platform on Bedrock.