The brief

A young fashion company had proven its idea on a no-code prototype. To take real customers and real money it needed a product it owned and could change: designers generate garments from a prompt or a sketch, turn them into production documents, sell them, and reach a factory that can make them.

We joined as the startup's engineering team and took it from prototype to public launch.

What we built

  • A clean move off the prototype. Every existing record restored, then one TypeScript application in place of the patchwork.
  • A conversational design studio with versions, image enhancement and upscaling, and spec sheets exported as PDF.
  • A 3D garment viewer in the browser, with augmented reality on Android phones.
  • An AI advisor that recommends whether to sell a design now, take pre-orders or test demand first.
  • A shop with card checkout and a product export for Shopify.
  • A portal for factories: inbox, quotes, company profile and team invitations.
  • Team workspaces, notifications, an invite-only beta gate, and an admin panel for users, partners, access, the blog and 3D models.

How

Before customers arrived, a security review led to private file storage with links that expire, so a design is only visible to the people meant to see it. What a factory partner may see is decided in one place in the code, not screen by screen, so confidentiality cannot slip through a new page.

AI usage runs on a metered free allowance, so a busy day cannot turn into a surprise bill. Consent, privacy and age checks were reviewed against GDPR before launch. Every change went to a separate test site first, with automated end-to-end tests in front of production.

What changed

  • Public launch about a week after the rebuild started.
  • The founders went from a prototype they could not change to a codebase, data and accounts they own outright.

Related