Hypothesis: if you define the basic bricks of your branding and UI design, an LLM can build (not code) the UI on the fly, based on the particular user request.

We are used to static layouts (you go to this page, then you can click on that button, ...). But seeing it from this POV feels like all apps are slow, and we are forced to learn how their operations work instead of them doing what we want.

Unfortunately, Vercel has widely tied the term "Generative UI" with "coding agents focused on frontend". But this Medium post explains it many times better than I could: https://medium.com/user-experience-design-1/an-introduction-...

I've started building an MVP for just one app, and it hasn't disappointed so far. It doesn't replace the entire app, but it does replace the main graphical parts the user interacts with. I've put a couple of mockups on the website linked above.

I'm planning to make it a generic framework for any app to be generative. Looking for thoughts, comments, ideas, or interest in trying it out as an early user (you will need some coding skills, it's not fully no-code).