Date: 2026-05-17

Audience: IT developers working with .NET + actor models + LLM integration

Sister page (Korean): Akka.NET에 LLM Agent 객체 만들기 — Akka.io Agent SDK를 액터 모델로 이식하기

Related memorizer notes (new window):

• Akka.NET Distributed Actor Model (Korean)

• Akka Agents Platform Overview (Korean)

• Akka Sample Projects (9 agents) (Korean)

• Akka Hello World Agent Tutorial (Korean)

• Akka Multi-Agent Planner (Korean)

• Actor-Agent RAG (Memorizer-v1, Akka.NET production case) (Korean)

TL;DR

Akka.io (on the JVM) ships an Agent component that bundles LLM calls + session memory + tool calls into a single first-class abstraction. Akka.NET does not yet provide this abstraction natively, but you can build the same shape using ReceiveActor + Become transitions + supervised child actors + an external LLM client. This article walks through the design and runnable samples.

Background — Why isn't there an Agent in Akka.NET?

Akka.io's Agent SDK wraps a Context Gathering → Reasoning → Action Taking → Progress Evaluation lifecycle into one abstract class (see Akka Agents Platform note). On the .NET side you have to build the same shape with explicit Become transitions + PersistentActor checkpoints.

1. Akka.io Agent in one line

Minimal form from the official Hello World tutorial:

@Component(id = "hello-world-agent")
public class HelloWorldAgent extends Agent {
    private static final String SYSTEM_MESSAGE = "You are a cheerful AI...";
    public Effect<String> greet(String userGreeting) {
        return effects()
            .systemMessage(SYSTEM_MESSAGE)
            .userMessage(userGreeting)
            .thenReply();
    }
}