Hi everyone! :wave: I just published a new article...

# koog-agentic-framework

[Vadim Briliantov]

Hi everyone! 👋
I just published a new article on Medium!

It explores how Domain Modeling approach can turn fragile prompt-based systems into trustworthy, type-safe AI workflows — using the Koog Framework as an example.

Thinking about structure and types can provide more control over the workflow then writing long textual essays in prompts

👉 https://medium.com/@vadim.briliantov/beyond-prompts-use-domain-models-to-rule-ai-agents-instead-89765faa6e46

[Balazs]

Hi @Vadim Briliantov, great article, do you have the example up on github?

[Eduardo Ruesta]

i think is this repo: https://github.com/JetBrains/koog/tree/develop/examples/devoxx-belgium-2025 @Balazs

[rcd27]

@Vadim Briliantov I have a huge prompt, which hit ~90%+ on evaluation tests(llm-rubric). How do you think, should it be split to different nodes? (your article is exactly about this, but I'm not sure in benefits I get, except parallelization maybe). Would be great if you take a look: https://github.com/rcd27/koog-self-research/blob/2e7f121400e09c8114ef4b9ef26c05cae[…]/rcd27/koogselfresearch/agent/strategy/subgraphLLMToolsBrief.kt