I wanted to explore the A2UI protocol using my new favorite tech/stack, so I did a heavily LLM-assisted implementation of it.

It's a very early release and more of an experimental nature. I have yet to use it in a sharp situation, but my goal is to do that. Will report back later with real world usage, and of course update and tweak the API so it's as fluid as possible.

And as it happened last time for me, I found another implementation (ex_a2ui) just as I published this. I think there is room for both as they have different use cases and APIs (this being server-side rendering, the other being the server protocol).

Maybe someone of you find this interesting still!

Thanks, Max