Inspiration

Aguamarina Store is a real handmade jewelry shop in Calabozo, Venezuela. Choosing jewelry online is often guesswork: the photo does not show how a piece looks on you, and color is hard to judge.

I wanted to turn the real catalog into a guided styling experience: discover your style, find a specific piece, and try it on from your phone without signing up.

What it does

Aguamarina AI is a styling adviser and virtual try-on experience built on the real catalog. A short conversation captures occasion, style and budget. YouCam analyzes skin tone, the adviser recommends real pieces from the catalog, and the shopper can try them on with their own photo or a model.

How I built it

HTML, CSS and vanilla JavaScript on the frontend; Node/Express and MongoDB on the backend; Cloudflare Pages and Pages Functions for hosting and AI proxies; Render for the API. The adviser uses a guided, rule-based conversation, while YouCam provides skin-tone analysis and virtual try-on.

The challenge

Virtual try-on quality depends on the product asset as much as on the API. I learned to curate pieces, prepare product images, and be transparent about the limits of a simulation.

What I learned

I learned to integrate visual AI into a real commerce flow, design a conversation that feels light, and ship a mobile-first experience without sign-up.

Built With

Share this project:

Updates

Submission history