Case study
Couple Suits
A shop for matching family outfits, plus a CMS with an AI design studio for the designer.
- Python
- FastAPI
- Jinja SSR
- PocketBase
- Gemini
- Cloudflare FLUX
- Pillow
- PWA

The problem
Matching outfits are usually sold as couple sets only. Families want the same look for parents, kids and siblings, and the designer needed a faster way from sketch to product.
What I built
- Server-rendered storefront with a family set builder (roles → cuts → sizes), cart re-priced on the server, family-set discounts, vouchers and WhatsApp ordering.
- Admin PWA: products, stock, orders with WhatsApp templates, homepage builder, journal, lookbook, reviews, cookie-free analytics, push on new orders.
- AI design studio: hand sketch + detail photos → Gemini brief → FLUX design options → flat drawings → the whole family's set → tech pack with graded size charts → PDF → draft product.
Highlights



Try it
Build a family set
Pick roles and sizes; the family-set discount kicks in at 3 and 5 people, as in the shop.
How it works
- 1Sketch + details
- 2Gemini brief
- 3FLUX options
- 4Flats + family board
- 5Tech pack + size charts
- 6Draft product
Engineering notes
- Size charts are graded from standard Asian tables using the AI's fit notes, not free-form AI numbers.
- Images are processed into WebP sizes; the admin keeps private studio files outside the public media folder.
How I build
Designed, built, tested and deployed by me with an AI coding assistant (Claude Code): I set the requirements, review every change, test on scratch copies with fake data, and run it in production.