case://study-001
iOS
Project 001

TryIt

A native iPhone app that composites clothing onto the user's selfie via Gemini 3 Flash Image. Three input modes (paste a product URL, share a screenshot via the system Share Sheet, or point the camera at a tag), a two-stage AI pipeline (a cheap Gemini 2.5 Flash validation pass before the expensive image-gen call), an 8-screen onboarding funnel tuned for conversion, RevenueCat-managed subscriptions, a Lock-Screen widget showing the latest result, and a privacy posture Apple's review team has on file: photos never leave the device beyond the ephemeral AI request.

Download · App StorePrivate repo · Source on request
App Store
Status
App + Widget + Share
Targets
~9k Swift
Code
iOS
Category
The Problem

What Wasn't Working

Online clothing returns are a $24B problem because people can't tell how a garment will look on them. Existing try-on tools either need a brand-specific integration or upload your face to a marketing CRM.

The Solution

How I Fixed It

A single-tap iPhone app: drop in a product link or screenshot, the AI extracts the garment and composites it onto a reusable profile selfie. No account, no third-party upload beyond the ephemeral AI request. Three discovery surfaces (link / share / camera) cover every way a user encounters clothing in the wild.

Stack

Technologies Used

SwiftUI
iOS 18
Gemini 3
RevenueCat
WidgetKit
Share Extension
XcodeGen
Results

Key Outcomes

Shipped to the App Store (review-cleared on face-data + AVCaptureSession bug fix)
Two-stage AI pipeline cuts cost on invalid inputs before the expensive generation step
Widget + Share Extension turn the app into an OS-wide capability
On-device-only photo storage — privacy posture documented to Apple Review

Want something like this?

Let's build it. I ship fast and I ship clean.