Back in 2019, Microsoft Open-Sourced Dapr, a new execution year for the easier operation of microprocessed applications. At that time, no one was still talking about AI agents, but as it turns out, Dapr had some of the fundamental building blocks to support AI agents integrated from the beginning. This is because one of the main features of Dapr is a concept of virtual actresseswhich can receive and process messages, regardless of all other actors in the system.
Today, the Dapr team starts the Dapr agents, the planning of the developers is manufacturing AI agents by providing them with many building blocks to do so.
“Agents are a very good use for Dapr,” explained co-creator Dapr and conservator Yaron Schneider. “From a technical point of view, you could use the actors as a very light way to execute these agents and really be able to run them on a scale with the situation-and be efficient. All this is great, but then, there is still a lot of operational logic that you have to write. But there is still a lot of work they have to do to really write the logic of the agent on the other hand.
Dapr agents came from GrooveA popular open source project expanded by Dapr for this case of use of AI Agent. Speaking to project conservationists, including Microsoft AI Roberto Rodriguez, the two teams decided to bring the project under the Dapr umbrella to ensure the continuation of the new framework agent.
“In many ways we see Agentic Systems and the whole terminology around it as another term for ‘distributed systems,’ said co-creator Dapr and conservator Mark Fussell. ‘[…] Instead of calling them small businesses, you can call them agents now, mainly because you can put large linguistic models between them all. ”
To effectively coordinate these factors, you need a orchestration machine and state status, the team argues – exactly Dapr delivers. This is partly due to the fact that Dapr actors are intended to be extremely effective and capable of returning within milliseconds when a message comes (and ending, with their state being maintained when their job is done).
Right now, Dapr agents can talk to most of the popular model providers from the box. These include AWS Bedrock, Openai, Anthropic, Mistral and Hugging Face. Support for local LLMS will arrive very soon.
In addition to interacting with these models, as Dapr agents are expanding the existing DAPR framework, developers are also able to define a list of tools that the agent can use to accomplish a given task.
Currently, DAPR agents soon support Python, with the start of the .net support. Java, JavaScript and Go will soon follow.
