Glowverse App:AI beauty: virtual try-on, skin analysis

Glowverse =an ai/ar-powered beauty shopping app offering virtual try-on, skin analysis, and personalized product recommendations with seamless e-commerce.

Comment

Inspiration

What it does

How we built it

Challenges we ran into

Project Name

  • Glowverse Problem

  • Online beauty shopping is guessing without seeing. Customers can’t reliably visualize shades or finishes, struggle to understand skin needs, and drop off at checkout due to uncertainty. Solution

  • A cross‑platform mobile app that lets people virtually try makeup in real time, run quick skin analysis, and get personalized product recommendations—backed by a reliable e‑commerce flow and privacy‑aware analytics. What It Does

  • Real‑time AR try‑on for lipstick, eyeshadow, eyeliner, blush, and foundation.

  • Quick skin analysis that highlights concerns and suggests routines.

  • Smart search and filters, wishlists, cart, and streamlined checkout.

  • Shareable screenshots and referral flows to bring in friends. Why It’s Different

  • Fast AR pipeline with device‑aware performance tuning (no perceptible lag).

  • Offline‑first request queue so cart and actions work even with spotty network.

  • Thoughtful UX: accessible components, haptic micro‑interactions, clear guidance.

  • Privacy-first analytics (no PII; meaningful funnel metrics only). Core Features

  • AR Try‑On: Live overlays, capture, intensity control, comparison.

  • AI Skin Analysis: Consent flow, result highlights, and product mapping.

  • E‑Commerce: Full catalog, cart, multi‑step checkout (Stripe, Apple/Google Pay).

  • Discovery: Fuzzy search, category filters, promo codes, guides.

  • Profile & History: Orders, addresses, saved looks, and analysis history. Tech Stack

  • Frontend: React Native 0.81.5, Expo SDK 54, TypeScript.

  • AR: Vision Camera pipeline with a native bridge for AR SDK integration.

  • State/Data: React Context + TanStack Query, Axios interceptors.

  • Payments & Analytics: Stripe React Native; privacy‑aware event tracking.

  • Quality: Jest unit tests, Detox E2E suites, Sentry for crash/perf tracking. Architecture

  • Modular screens and 76+ reusable components organized by domain.

  • Typed API layer with retries, deduplication, and health monitoring.

  • Design system with light/dark themes and consistent tokens.

  • Navigation via React Navigation with deep links and universal links. Implementation Highlights

  • Frame processor + native bridge keeps AR responsive on mid‑range devices.

  • Offline queue replays failed network actions automatically on reconnect.

  • Product image preloading and list virtualization keep scrolling smooth.

  • Error boundaries and focused fallback UI for a resilient feel. User Experience

  • “Try Before You Buy” flow reduces uncertainty at the moment of choice.

  • Clear callouts for shade match, finish, and intensity adjustments.

  • Skin analysis result cards explain findings in plain language, not jargon.

  • Checkout is one pass: review, confirm, pay—no detours or surprises. Impact

  • Fewer returns and higher conversion by removing shade guesswork.

  • Better engagement via shareable looks and referral perks.

  • Faster browsing and smoother sessions on both iOS and Android. Demo Plan

  • Open the app, scan the face, apply a lipstick shade, adjust intensity.

  • Run a quick skin analysis, view top concerns and recommended products.

  • Add items to cart, apply a promo, and complete a test checkout.

  • Capture and share a before/after screenshot. Security & Privacy

  • No PII in analytics; sanitized event properties only.

  • Secure token storage; crash/performance monitoring without sensitive data.

  • Consistent permission prompts (camera, photos) with clear reasons. What We Learned

  • AR feels magical only when performance is invisible—frame timing matters.

  • Simple, honest copy beats cleverness in analysis results and promos.

  • Offline‑first is worth the investment; it makes the app feel reliable.

  • Small accessibility details (touch targets, focus states) lift the whole experience. Roadmap

  • Expand payments with webhook‑driven order state and edge‑case coverage.

  • Broaden AR categories and add side‑by‑side comparison modes.

  • More robust skin concern tracking and longitudinal progress views.

  • CI‑driven store submissions and release automation. This is a practical, working app built for real users: quick to pick up, enjoyable to use, and focused on solving a clear problem—bringing confidence to beauty shopping through try‑on and insight.

    Accomplishments that we're proud of

What we learned

What's next for Glowverse App:AI beauty: virtual try-on, skin analysis

Built With

  • and
  • and-a-custom-native-bridge.-analytics-and-crash-reporting-use-a-privacy-aware-wrapper-and-sentry;-payments-are-powered-by-stripe-for-cards
  • and-artillery
  • and-asyncstorage
  • and-configuration-uses-expo-build-properties
  • and-eas-build/submit.-cloud-and-infrastructure-span-aws-ecs/rds/s3-(managed-via-terraform-and-kubernetes-hpa)
  • and-expo-sdk-54.-the-ui/ux-layer-uses-react-native-paper
  • and-google-pay;-notifications-use-expo-notifications-and-firebase.-media-and-device-utilities-include-expo-image-picker
  • and-health-checks.-ar-and-camera-features-leverage-vision-camera
  • and-lottie
  • and-netinfo.-the-backend-runs-on-node.js-18-and-express-4.21-with-prisma-6.3
  • and-the-web-using-react-native-0.81.5
  • and-zod.-testing-covers-jest
  • android
  • apple-pay
  • babel
  • blur
  • built-with-typescript-and-javascript-across-ios
  • cloudinary
  • deduplication
  • detox-e2e
  • docker
  • expo-camera
  • file-system
  • gesture-handler
  • linear-gradient
  • mmkv
  • plus-render-and-railway
  • postgresql
  • react-19.1
  • reanimated
  • redis-(ioredis)
  • secure-store
  • sharp
  • supertest
  • tanstack-query
  • testing-library-react-native
  • typescript-5.9
  • while-networking-is-handled-by-axios-interceptors-for-retry/backoff
  • winston
  • with-ci/cd-on-github-actions
  • with-navigation-via-react-navigation-7-supporting-deep-linking-and-universal-links.-state-and-data-are-managed-with-react-context
Share this project:

Updates