AI-powered selfie transformation

SwagCity — an AI-powered iOS app that turns a selfie into a stylised portrait — built solo, in seven days: concept to App Store submission.

I worked across brand, product strategy, UX, iOS build, backend and AI infrastructure the whole way, using Claude Code and ChatGPT as working partners throughout the build rather than adding AI in after the fact.

Solo build · 7 days, concept to submission · iOS consumer app

Project snapshot

The challenge

Every layer of a real consumer app, one person, one week.

SwagCity needed to look and feel like a trustworthy, polished consumer app — not a weekend experiment — on a timeline with no separate design phase, build phase or compliance phase. Product decisions, interface design, working code, a real payment system and Apple's App Store requirements all had to move together, from day one.

Some of the core features are

• Mobile App for consumers
• Backend Dashboard App for admins to manage styles/tokens/subscriptions

Why it was complex

Speed and craft aren't usually the same week

Most solo AI-app builds either move fast and skip the details App Review checks for, or move carefully and take months. SwagCity had to do both in the same sprint — a fast, low-friction core loop, and a fully compliant, production-grade credit and payment system.

Real money, from day one

Because the app charges for AI generations, the credit economy had to be right the first time — credits are minted server-side only, never trusted from the client, with no soft-launch period to quietly patch a broken payment flow later.

No separate research phase to fall back on

With the whole build compressed into a week, user feedback had to be gathered and acted on inside the build itself — not after a public launch.

My role

Concept to submission, one system at a time.

I owned every layer — product, design, iOS build, backend, payments, testing and the App Store submission — working solo across Figma, Claude Code, Expo, Supabase and Replicate.

Day 1 — Concept and system orchestration

Used Claude Code and ChatGPT to move from idea to a working plan fast — mapping the roadmap, the core user journey (snap a selfie, pick a style, swagify and share) and a phased build plan, and setting up how the different AI tools would work together across the rest of the build.

roadmap.md users.md phases and mvp.md wireframes brand elements

Why wireframe artefact when i can directly do prototype in claude code?

Answer is simple, i use my own low cost wireframe system to run with claude skill which is trained in creating wireframes very cost effectively.

Saving tokens :)

Day 2 — Building the app

Built the core iOS app in Expo React Native — navigation, camera and photo-library capture, and the main screens.

Expo scaffolding Iterative Builds

Day 3 — Theme and backend foundations

Built the visual theme and the Supabase backend — style and category management, a server-authoritative credit economy, and account management.

Develop theme Schema planning and supabase Batch style generation using chatgpt and openArt

Day 4 — Payments

Built the payment plans and payment gateway behind a paywall — consumable credit packs and a weekly/yearly Pro subscription via RevenueCat and Apple's in-app purchases.

Revenue cat config and setup App store product setup

Day 5 — TestFlight

Built and released the app to TestFlight for testing.

Test 1 - found many bugs and improvements

3 styles, basic paywall

Test 2 - improvements and bug fixes

60 styles, advanced paywall

Day 6 — Testing, security, compliance and submission

Ran user testing with a group of pilot users among friends and family, gathered their feedback and used it to polish the user journey; ran a security audit; worked through Apple's App Store compliance requirements — account deletion, Sign in with Apple, privacy labels; and submitted the app for App Store review.

Product Strategy iOS Development Backend & AI Pipeline User Testing

User testing

Multiple users

• 3 users on test flight, concurrent sessions, generation privacy
• RSL policies and fixing security bugs
• Fix genZ styling gaps

Impact at a glance

Concept to App Store submission in 7 days

Every layer — product, UX, iOS build, backend, AI pipeline and payments — designed and built solo, one system at a time.

Real user feedback folded into the build itself

Pilot testing with friends and family on day 6 directly reshaped the user journey before submission, not after launch.

Currently awaiting App Store review

The app has been submitted to Apple and is awaiting review. Fixing privacy policy notes.

Swagcity Website