Skin2Style AI

• 

  Real YouCam Clothes V3 result using an official sample person and garment reference.

• 

  Consented YouCam Skin Analysis supports color guidance. It is not medical advice and never rates attractiveness.

• 

  Responsive mobile styling with readable recommendations, clear navigation, and touch-friendly controls.

• 

  Anonymous journey, product, conversion, and API-unit insights—without storing photos or body measurements.

• 

  Real YouCam AI Shoes result, continuing from the clothing transformation with virtual footwear.

• 

  Choose an occasion to tailor the styling goal, outfit formula, palette, and product recommendations.

• 

  Private AI styling with body guidance, skin analysis, virtual try-on, nails, makeup, and retail insights.

Inspiration

Beauty and fashion decisions rarely happen separately. People often choose skincare, makeup, and clothing across different platforms, without a clear way to understand how those choices work together for a specific occasion.

Skin2Style AI was created to connect these decisions in one guided retail experience. Instead of showing users isolated skin scores or a single virtual try-on result, it helps them move from understanding their visible skin condition to choosing and visualizing an appropriate, shoppable outfit.

What it does

Skin2Style AI guides users through a simple five-step journey:

1. Select an occasion, such as an interview, date, wedding, or everyday shopping.
2. Upload a selfie for YouCam Skin AI analysis.
3. Receive responsible, non-medical skin observations.
4. Explore clothing recommendations ranked by occasion, budget, color, size, inventory, and available skin-analysis signals.
5. Upload a full-body photo and use YouCam Apparel Virtual Try-On to preview the selected item.

The final screen combines the virtual try-on result, recommendation score, product information, and explainable reasons into a shoppable look.

A retail dashboard also tracks the user funnel, popular products, virtual try-on activity, and YouCam API unit usage.

How we built it

The prototype uses a responsive HTML, CSS, and JavaScript frontend with a Python FastAPI backend.

The backend keeps the YouCam API key secure and handles image uploads, asynchronous task creation, status polling, result processing, caching, and API-unit tracking.

Our recommendation engine uses a transparent rule-based scoring model:

• 30% occasion compatibility
• 25% user preferences
• 20% color coordination
• 15% budget and inventory
• 10% virtual try-on feedback

This allows every recommendation to include a reason based on real product attributes and available API results.

Challenges we ran into

The main challenges were managing asynchronous image-processing jobs, controlling API-unit consumption, normalizing API response formats, and preparing suitable user and garment images for reliable virtual try-on results.

We also needed to design skin-related language carefully. Skin2Style AI presents visible cosmetic observations without making medical diagnoses or negative judgments about appearance.

Accomplishments that we're proud of

We designed Skin AI and Apparel VTO as one connected shopping journey instead of two unrelated API demonstrations.

We also created an explainable recommendation engine, server-side API-key protection, result caching, API-unit accounting, and a retailer-facing conversion dashboard.

What we learned

We learned that successful AI retail experiences depend on more than model output. Image quality, product metadata, error recovery, understandable recommendations, privacy communication, and API cost controls are equally important.

What's next for Skin2Style AI

Next, we plan to expand the licensed product catalog, add side-by-side comparison for multiple virtual try-on results, improve image-quality validation, strengthen consent and deletion controls, and deploy the complete experience for public testing.

Built With

• analysis
• api
• apparel
• css3
• fastapi
• html5
• javascript
• python
• rest
• skin
• try-on
• virtual
• youcam