Case study — Fashion e-commerce
Rosa Boutique.
A boutique from Erbil and Dubai that speaks Kurdish first.
Try it live
The real thing, right here.
The brief.
Rosa sold abayas and evening dresses only through Instagram. Every order was a DM, every price a question. The shop needed a real home that still felt like Rosa, and that spoke to women in the language they actually use.
The dig.
Kurdish (Sorani) is the default, with Arabic and English one tap away. Every photo and video is Rosa’s own, pulled from their account. The design is a fashion magazine: a Bodoni headline, a clean Jost body, blush and wine from their logo, and not one rounded corner.
The shine.
A shop and a staff dashboard (admin, orders, inventory) in one build. Discount codes are named after things the customers love: Newroz and Hawler. The opening gift (perfume for the first 50 buyers) lives right on the site.
The system
How it looks, and why.
Blush and wine, a magazine headline, sharp edges. Quiet luxury that still sells.
#EAD3CFSoft fields, highlights#C98C7AAccent, the gold of the brand#6E1E2EDeep accent, buttons#1A0B10Text, a wine-black#F5F3F3Behind every product photo#E4DEDEHairlines- Every radius is 0. Sharp is the brand.
- No shadows. Depth comes from photos, not effects.
- Each language gets its own fonts, not a fallback.
- Next.js 16
- React 19
- next-intl
- GSAP
- Go API (next)
Worth stealing
Ideas from this one.
Kurdish is the default
Most shops treat Kurdish as an extra. Here it is the first page you see.
Real over stock
Every image and reel is Rosa’s own. The site feels like their Instagram, only easier to buy from.
Local codes
NEWROZ15 and HAWLER aren’t random codes. They are holidays and a city the customers love.
Grey behind products
Product photos sit on #F5F3F3, so every dress reads the same way, whatever room it was shot in.
Seen enough? Go and use it.
rosa-ashy-alpha.vercel.app