Glowra: AI Skin Analysis

Inspiration

Good skincare advice is expensive, generic, and hard to stick with. We wanted an AI skincare coach in everyone's pocket — one that looks at your actual skin, explains what's going on in plain language, and adapts as you improve. And we wanted the business model done right from day one, not bolted on later.

What it does

Glowra turns your phone camera into an AI skincare coach:

• Scan your face and get a dermatologist-style cosmetic analysis in seconds (hydration, texture, evenness, and more).
• Beauty Chat AI — a 24/7 skincare assistant that answers your questions using your latest scan as context.
• A routine that adapts to you and updates as your results change.
• Product recommendations matched to your skin profile.
• Progress tracking with glow streaks and a visual timeline so you can watch your glow-up.
• Glowra Premium unlocks unlimited scans, full Beauty Chat, and complete analysis history — an auto-renewable subscription.

How we built it

• React Native + Expo (TypeScript) for one cross-platform iOS/Android codebase, shipped with EAS Build & Submit.
• RevenueCat (react-native-purchases) for the entire subscription layer — entitlements, offerings/packages, the paywall, and cross-platform purchase + restore — so we never hand-roll StoreKit or Play Billing.
• Supabase for auth (Sign in with Apple), Postgres, secure per-user image storage with row-level security, and edge functions.
• AI: computer-vision skin analysis plus an LLM-powered Beauty Chat (Azure OpenAI) that grounds every answer in the user's most recent scan.

Challenges we ran into

• Shipping a first auto-renewable subscription: the subscription, its subscription group, and the app build all have to be submitted to Apple together — a subtle multi-item flow that's easy to get wrong.
• Cross-platform entitlement logic that resolves the correct current offering on each store.
• Handling face images privately and securely (per-user storage paths + RLS).

Accomplishments that we're proud of

• A real, monetizable product — not a prototype — with RevenueCat wired end to end and verified against the live offerings endpoint.
• Shipped to the App Store (in review) and Google Play closed testing.
• An AI experience that feels personal because every response is grounded in your actual skin scan.

What we learned

• RevenueCat's entitlement model turns "does this user have premium?" into a single reliable check and kept our paywall logic trivial across platforms.
• Clear paywall entry points matter as much as the purchase code — reviewers (and users) have to actually reach the offer.

What's next for Glowra

• An annual plan and free-trial intro offer, tuned with RevenueCat Experiments.
• Shareable glow-up cards to create viral loops.
• Deeper progress analytics and dermatologist-reviewed routines.

Built With

• android
• azure-openai
• eas
• expo.io
• ios
• postgresql
• react-native
• revenuecat
• supabase
• typescript