FitCheck AI Stylist

Inspiration

Getting dressed is easy. Knowing whether the outfit works for where you’re going is the harder part. I wanted a quick second opinion without having to post a photo online or ask someone every time. That became FitCheck: take a photo, choose the occasion, and get useful feedback. The focus is on the clothes and how they work together—not on judging the person wearing them.

What it does

FitCheck gives your outfit a score and explains it. It looks at things like color, fit, shoes, accessories, and whether the look suits the occasion. It also suggests changes, estimates the value of visible pieces, and lets you compare outfits. You can save your results or share a scorecard with your photo, score, and a QR code for downloading the app. The price estimates are just that—estimates, not confirmed prices.

How we built it

I built the app with SwiftUI for iPhone and iPad. Photos go through a Cloudflare Workers backend to OpenAI for analysis, and the response becomes the scorecard you see in the app. RevenueCat handles the Pro paywall and purchase flow, alongside Apple’s StoreKit verification. Users give permission before a photo is sent for analysis, and their saved outfit history stays on their device.

Challenges we ran into

Some of the most frustrating problems were in the details. Earlier scorecard exports cropped photos, cut off descriptions, or left out the outfit-value estimate. A result could look fine inside the app and still export badly. Purchases brought another round of troubleshooting. The sandbox account initially refused purchases, so getting through a successful checkout on a real iPhone—and seeing Pro unlock—was an important test.

Accomplishments that we're proud of

FitCheck is a working app, not just a set of mockups. It can analyze an outfit, explain its feedback, and turn the result into something worth saving or sharing. I’m also happy with the iPad support and the RevenueCat integration. The purchase flow has now been tested successfully on a physical iPhone.

What we learned

The score alone isn’t enough. People need to understand why they got it and what they can actually change. Testing also has to include the less exciting parts: exported images, long descriptions, missing clothing details, and purchase errors. Those details can make an otherwise useful app frustrating.

What's next for FitCheck AI Stylist

I want to improve the consistency of the advice and make uncertainty clearer, especially around brands and price estimates.

There’s also more purchase-restoration and subscription testing to finish. After that, the priority is getting feedback from people using FitCheck with their own clothes and occasions, then improving what they find useful.

Built With

• cloudflare-workers
• ios
• openai-api
• revenuecat
• revenuecatui
• storekit-2
• swift
• swiftui
• typescript
• xcode