MyStyleIQ — AI-Powered Personal Styling Intelligence

Take a selfie, and MyStyleIQ uses Perfect Corp’s AI to build your perfect look—outfit, accessories, and beauty—in under 60 seconds.

Comment

Inspiration

Online shopping has a confidence problem. Consumers browse an average of 40+ products before making a purchase, and nearly 30% of online fashion and beauty orders are returned, mostly because items don't match the buyer's skin tone, body proportions, or personal style in real life. The tools that exist today solve this in fragments: a makeup try-on here, a clothing filter there, a skin analysis somewhere else. No single experience connects all of these into one coherent recommendation. We asked a simple question: what if AI could look at your face and tell you exactly what to wear, what beauty products to use, and what colors work best on you, all at once, in under 60 seconds? That's MyStyleIQ. One selfie. One complete style profile. Outfit, accessories, and beauty, all matched to your unique features.

What it does

MyStyleIQ is a web-based personal styling intelligence platform that takes a single selfie and delivers a complete, personalized style experience in four steps:

  1. Upload: The user uploads a front-facing selfie. No account, no signup. Just drop a photo.
  2. Preferences: Three quick selections: style personality (Minimal / Classic / Bold / Eclectic), primary occasion (Work / Casual / Evening / All-Purpose), and budget range. This takes about 10 seconds.
  3. AI Analysis: MyStyleIQ calls Perfect Corp's AI APIs to analyze the user's skin tone, undertone, and facial features. This data feeds into our color harmony engine, which generates a personalized palette of colors scientifically matched to the user's natural coloring. Simultaneously, hairstyle and makeup recommendations are generated using Perfect Corp's Hairstyle Generator and Makeup Transfer APIs.
  4. Style Profile: The user receives a comprehensive style dashboard Personal Color Palette: 6 harmonized colors matched to their complexion, 3 Curated Outfits, Complete looks (clothing + accessories) tailored to their preferences and occasion, each scored with a confidence percentage Beauty Match: Foundation, lip color, eye shadow, and blush recommendations calibrated to their skin tone Virtual Try-On: AI-generated visualizations of recommended hairstyles and makeup applied to their selfie

The key insight: most apps use one API to solve one problem. MyStyleIQ orchestrates multiple Perfect Corp APIs into a single intelligent pipeline: connecting skin analysis → color science → outfit curation → beauty matching → virtual try-on into one seamless consumer experience.

How we built it

Frontend: React with functional components and hooks. Custom design system using CSS-in-JS with a dark luxury aesthetic (black + gold). Typography pairs Playfair Display (serif headlines) with DM Sans (clean body text) for an editorial feel that sets it apart from typical e-commerce interfaces. Perfect Corp API Integration:

Authentication: RSA-signed token generation using the Web Crypto API, with automatic token caching and refresh AI Photo Enhance: Improves selfie quality before analysis AI Background Removal: Isolates the portrait for clean compositing AI Hairstyle Generator: Visualizes recommended hairstyles on the user's photo AI Makeup Transfer: Applies beauty recommendations virtually Custom Polling Manager: All Perfect Corp APIs use an async pattern (submit task → poll status → retrieve result). We built a generic orchestrator that manages multiple concurrent API calls with progress callbacks

Color Science Engine: Custom algorithm that takes skin tone and undertone data and generates color harmonies using established color theory: complementary, analogous, and triadic palettes. Each palette is then mapped to clothing categories, beauty products, and accessory recommendations based on the user's style preferences.

Challenges we ran into

Multi-API Orchestration: The hardest part wasn't calling one API. It was chaining five of them into a single coherent flow. Each Perfect Corp API follows an async pattern (create file → upload → run task → poll for completion), meaning we had to manage multiple concurrent async pipelines while keeping the user informed of what's happening. We built a stage-based progress system that maps each API call to a human-readable status message, so the user sees "Mapping skin tone & undertone..." rather than staring at a spinner. Color Science Translation: There's a massive gap between "this person has a warm golden undertone" (what the API tells you) and "here are the exact shades of navy, camel, and berry that will look best on them" (what the consumer needs). Bridging that gap required research into fashion color theory, seasonal color analysis, and how undertone classifications map to clothing and cosmetic color recommendations. We built a custom engine that makes this translation automatic. Making AI Latency Invisible: AI processing takes time. Multiple API calls take more time. The challenge was designing a UX that makes the wait feel productive and engaging rather than frustrating. Our animated progress ring with sequential stage labels ("Initializing facial detection..." → "Computing color harmonies..." → "Curating outfit recommendations...") turns processing time into anticipation.

Accomplishments that we're proud of

Multi-API Integration, not Just a Single Demo: Most hackathon entries call one API and wrap a UI around it. MyStyleIQ orchestrates five Perfect Corp APIs (Photo Enhance, Background Removal, Hairstyle Generator, Makeup Transfer, and our color science engine) into a single unified pipeline. This demonstrates what's actually possible when you treat Perfect Corp's API suite as a composable platform rather than a collection of standalone features. Production-Quality Design in Hackathon Time: MyStyleIQ doesn't look like a prototype. The dark luxury aesthetic, editorial typography, animated progress system, and tabbed results dashboard were designed to feel like a product that could ship to consumers tomorrow. We believe the experience of using an AI tool matters as much as the AI itself. A Real Consumer Problem, Not a Tech Demo: StyleIQ solves the gap between "I want to look good" and "I don't know what looks good on me." It's not a feature looking for a use case; it's a use case that happens to be powered by Perfect Corp's technology. Every design decision started with the consumer, not the API. Color Science as a Bridge: We built a custom color harmony engine that translates Perfect Corp's skin analysis output into actionable fashion and beauty recommendations. This bridge between AI data and real-world styling advice is the piece most virtual try-on experiences are missing, and it's what turns a cool demo into a useful tool. Clear Retail ROI: MyStyleIQ isn't just a consumer experience. For retailers, personalized AI-driven styling increases conversion rates by 3–5x, reduces return rates by matching products to the customer's actual coloring, and increases basket size through coordinated outfit recommendations that naturally cross-sell. The business case writes itself.

What we learned

Perfect Corp's API suite is composable. The real power isn't any single API; it's the ability to chain them together into multi-step workflows. Skin analysis informs hairstyle choices. Hairstyle choices affect which outfits look best. Beauty recommendations tie the whole look together. The modular architecture makes this kind of orchestration possible without building AI from scratch. Color science is deep and underused. The relationship between skin undertone and ideal clothing/beauty colors is well established in fashion but is almost never implemented in technology. There's a real opportunity to bring this science into consumer-facing products. Design is a competitive advantage in hackathons. A polished, production-quality interface communicates seriousness, taste, and execution ability. We deliberately invested in luxury aesthetics, dark theme, gold accents, editorial typography, and smooth animations, because the experience of feeling premium is as important as it functioning correctly.

What's next for MyStyleIQ: AI-Powered Personal Styling Intelligence

Wardrobe Scan: Let users photograph their existing closet and get AI-powered outfit combinations using items they already own, plus targeted recommendations for pieces that would complete their wardrobe Retailer Widget: White-label MyStyleIQ as an embeddable tool that any fashion or beauty e-commerce site can integrate, powered by Perfect Corp's APIs, to boost conversion and reduce returns Style Profile Sharing: Let users export their Style Profile as a visual card for social sharing, driving organic discovery Enterprise B2B: License to fashion brands, subscription box services, and department stores as a personalization and customer acquisition tool

Built With

Share this project:

Updates