It is true that LLMs just don 'get' UX. It is a lot of trial, error, hallucination, fixing to get to a working state for even simple pages. If there are no specifics, it can generate a nice looking page to impress people, but add functionality, buttons, layouts, login, list, image tags and suddenly it goes into never ending loop of hallucinations. Claude still does better than other.

This is funny in a slightly painful way. We made the web more and more visual for humans, and now we need tools to strip it back down so agents can understand it.. Kind of feels like reinventing RSS for LLMs. Nice project,this solves a very real annoyance

[flagged]

[dead]

[dead]