Velour: Outfits & AI Try-On

See yourself in styles you've never tried: complete looks, and an AI try-on from one photo of you.

Comment

Show your full potential

Most people know there's a better-dressed version of themselves. They just don't have time to work out what it looks like, or they're stuck in one style and unsure whether anything else would suit them.

Velour is for two kinds of people:

  • People who want to look good without researching it. Velour gives you complete outfits that are already put together, and every piece links to similar items from real retailers.
  • People with a set style who want to try something new. You can browse looks by style and by occasion, then see them on yourself before you spend money or courage on them.

Velour isn't a closet organiser. It shows you what else you could wear.

How it works

  1. Swipe to set your taste. On first launch you swipe through a few looks, right to like and left to pass, and your feed adapts.
  2. Browse complete looks. Filter by style (Casual, Minimal, Classic, Chic, Streetwear, Boho and more) or by plans (Everyday, Work, Date, Going Out, Travel). Discover also suggests looks for today's weather.
  3. Open any look. Velour shows when to wear it, what weather it suits and every piece in the outfit. Each piece has "Items to Shop" with similar products from real retailers.
  4. Try it on. Add one front-facing, full-body photo. Velour checks it first (full body, front-facing, good lighting, just you), then renders any outfit on you. Your try-ons are saved in your Closet, so you can compare completely different styles side by side.
  5. Get more try-ons. Each try-on uses one credit.

Monetization, powered by RevenueCat

Velour sells one-time credit packs, not a subscription, and credits never expire. People try looks in bursts, before a date, a trip or an interview, so paying per try-on fits better than a monthly fee.

  • Google Play packs: 5 try-ons for $3.99, 20 for $9.99, 50 for $19.99 (Most Popular) and 100 for $34.99.
  • Free credits: new accounts start with 5 free try-ons, every 5th daily check-in adds one more, and an invite gives both friends 3.

RevenueCat runs the entire Android purchase flow:

  • In the app: purchases_flutter shows the Google Play packs, handles already-owned purchases and restores purchases.
  • On the server: a RevenueCat webhook sends each purchase to our FastAPI backend. Every transaction is credited exactly once, recorded against its Google Play order ID, and taken back exactly once if refunded.
  • Recovering interrupted purchases: Google Play real-time developer notifications reach us through RevenueCat, so a purchase is still credited if the app closes during payment.
  • Fail-closed: the balance only ever shows credits the server has confirmed.

How we built it

  • App: Flutter, with Riverpod and GoRouter.
  • Backend: FastAPI on Azure Container Apps, with Supabase (Postgres and pgvector).
  • Catalog: a curated library of AI-rendered editorial looks, tagged by style, occasion and weather.
  • Try-on: an image-generation model renders the chosen outfit onto the user's own photo, after the server has checked the photo.
  • Safety: every try-on result has a Report button. Users can delete their photo, or their whole account, in the app or on the web.

All try-on results and catalog images are AI-generated, and the app labels them that way. The model in our demo video is AI-generated too.

What's next

  • Bring Velour to iOS.
  • Send "Items to Shop" links straight to partner retailers through affiliate programs.
  • Add a "style journey" that suggests your next step outside your comfort zone, based on the try-ons you save.

Built With

  • azure-container-apps
  • azure-openai
  • dart
  • fastapi
  • flutter
  • google-play-billing
  • pgvector
  • postgresql
  • python
  • revenuecat
  • riverpod
  • supabase
Share this project:

Updates

Submission history