Dapr’s microerssear performance already supports AI agents

Already in 2019, Microsoft Open daprA new time for facilitating the building of distributed applications based on a microservice. At the time, no one was talking about AI agents, but as it turns out, DAPR had some of the main building blocks of support for AI agents built from the beginning. This is because one of the main characteristics of DAPR is a virtual concept actorswho can receive and process messages regardless of all other participants in the system.

Today, the DAPR team is launching DAPR agents, assuming that it helps developers build AI agents, providing them with many of the building blocks to do so.

“Agents are a very good case for DAPR,” DAPR co-creator explained and Dapr supporter. “From a technical point of view, you could use the actors as a very easy way to manage these agents and really be able to manage them on a scale with the state-to be effective for resources. All this is great, but then there is still a lot of business logic to write. Its state and orchestration are only one part. And many people, they can choose the engine of the actor’s workflow or frame, but there is still a lot of work to do that they have to do to actually write the agent’s logic on the other side. There are many agents frameworks, but they do not have the same level of orchestration and state state that DAPR has. “

Dapr agents come from FlockA popular open source project that expanded DAPR for this case of AI agent. Talking to the project supporters, including Microsoft Ai Roberto Rodriguez researcher, the two teams decided to bring the project under the DAPR umbrella to ensure the continuity of the agent’s new framework.

“In many ways, we see agency systems and all terminology around this as another term for” distributed systems, “said the Creator and supporting DAPR Mark Fusel. “(…) Instead of calling them micro -season, you can now call them agents, most of all because you can put large language patterns among them all.”

In order to coordinate these agents effectively, you need an orchestration engine and state state, according to the team – that’s exactly what DAPR provides. This is partly because DAPR actors are intended to be extremely effective and to be able to rotate within milliseconds when a message enters (and off, their condition is saved when their work is done).

Currently, DAPR agents can talk to most popular models of models outside the box. These include the AWS base, Openai, Anthropic, Mistral and the hug. Support for Local LLMS will arrive soon.

In addition to interacting with these models, as DAPR agents extend the existing DAPR framework, developers also get the opportunity to define a list of tools that the agent can use to accomplish a task.

DAPR agents are currently supporting Python, starting .NET maintenance. Java, JavaScript and Go will follow soon.