Skip to main content

Product overview

Drape is a two-sided marketplace built around one hard-to-fake moment: seeing a real garment on your own body before you buy. Shoppers get try-on confidence; independent brands get a try-on-enabled storefront and a discovery channel.

"Try on the culture.""Check your fit before you commit." "See how clothing fits, drapes, and moves on your body before checkout."

(The internal tagline "Try it on without trying it on" describes the same idea; the live marketing voice is "Try on the culture.")

Two audiences, one marketplace

Brands

Shoppers

marketplace fee

Discover independent brands

Upload one body photo

Try on → fit / drape / movement

Checkout on Drape

Apply + get approved

Branded storefront on {slug}.drape.to

Catalog · try-on · analytics

In-store QR + website widget

The marketplace framing is deliberate: "Shoppers discover, try on, and checkout on Drape — your storefront is where the sale happens." Brands do not run a parallel checkout elsewhere; the sale closes on Drape and the brand gets a branded surface, Stripe Connect payouts, and analytics.

The three value pillars

Drape markets itself on three pillars, and each maps to something real in the architecture:

PillarWhat backs it
AI-powered try-onThe try-on pipeline — Fashn.ai + fal.ai, garment masking, per-body caching
Independent brandsThe brand offering — curated, culture-forward labels, not big-box retail
Your data stays privateThe security model — BIPA consent, private storage, server-side body-photo handling

Where each audience is served

  • Shoppers use the mobile app (iOS/Android, currently waitlist) and land via the marketing funnel at drape.to.
  • Brands apply and operate through the brand offering — a desktop portal (brand.drape.to) and a mobile brand portal, over the same backend.
  • Internal ops run through the admin tooling at admin.drape.to and the autonomous cron agents.

What makes it hard

  • Biometric data. A body photo is BIPA-regulated. Consent is enforced at the backend, storage is private, and the photo is never accepted from the client on a try-on request. See security.
  • AI that is right often, not always. The pipeline routes between models and post-processes garments, and can still "succeed" with a wrong render. See the try-on pipeline.
  • A platform mid-migration. The web tier runs on Postgres while the mobile app is still on Mongo — see backend evolution.
  • Solo operations. One founder runs product, backend, four surfaces, and ops — so the system leans on monitoring, rate limits, and audit docs to stay safe.