Hi HN!

Backstory - I've been writing backend code for years, and about a year ago I told my boss to shove it and went full solopreneur mode.

Problem is - many products need a nice web interface, and I'm a TERRIBLE designer. over the past few months, I kept feeding screenshots of website sections i found beautiful to v0/Claude, asking it to give me something similar for my own apps. This approach sometimes works, but often completely falls apart for more complex UI elements.

So, i built this chrome extension, which grabs the actual source code (heavily pre-processed and sanitized) and feeds that to the LLM as context, instead of just a screenshot - and it works way better! It lets you pick any element in a webpage (kinda like the chrome devtools element selector), and it then outputs a react + tailwind component of the selected element.

would love your feedback or any thoughts/ideas

Yuval