Introducing generative UI for the agentic era 🤯 no apps. no dashboards. every screen is generated the moment you ask, reshapes as you talk, then disappears. try now at

slop merchants really think prompting is better UX than clicking a button

The concept is interesting for sure. How it gets implemented is a different story.

imo adaptive UI feels better than prompting

yes but the ui looks so bad :(

Open to suggestions Andres

let me know if you wanna chat. do love the concept though, big on agentic dynamic UI, love this idea and think it's where things are headed.

SaaS is gone in the next few years, so you're early to the game. But this is the future. Ai will just generate anything we need and customize our computers in real time.

Definitely. Generative UI is a promising concept, soon it will become the standard in the space

How does this work? Does it take over iOS UI?

Not at all. All bundled within the app

Is it open source?

It is not. At least for now.

Is this llm based local or cloud

platform is model agnostic. supports both

SI eats UI

Great concept, will adopt

And how the fuck u wire a backend into this? lol

You can find out on our Beta launch 🙂

“Put these side by side” and suddenly I have a comparison I can click and edit, love it

Noted Max. It's an experimental feature, we'll be playing around with it and see what works best from a UX point of view

This could be Huge if it runs ultra fast and has optimized memory footprint. As a Linux OS I would actually switch. Would love to experiment. Any full OS plans or just app?

For now we'll be supporting Mac build and introduce Linux/Windows afterwards. Stay tuned!

So just an app? Any plans on making a complete OS where most of the OS UI does not exist but only exist temporarily. Would be the smallest OS where just need the AI model and architecture for interacting and rendering UI

I was thinking about that earlier today but tbh that's a completely different venture and direction. I believe someone will introduce this concept soon enough. Curious to what it would look like myself too.

Yea I'm already imagining an AI service where one setups up VM, chooses from 1000s of SKILLs .md files and just simple button to talk and communicate with AI and it builds the UI to your liking instead of being locked down to specific skills or specific fixed UI waiting on update

Imagine that as an Android OS Distro, boot up choose skills no apps just button to communicate build your own UI for own custom operating system

oss?

Are there repeated interfaces and learning patterns? Getting UI into long-term memory is beneficial, I believe.

I think this is where it goes. Only without macOS underneath. Not a prompt floating on a desktop: this just is your computer. You ask, a screen appears, you are done, it goes away. Slightly awkward to say for someone who spends his days building apps.

can we get a real demo and not a polished motion vid

this is jev/decision api for sure

@daradoescode

Okay better than Siri