SkinIQ: AI Skin Consultant & E-commerce

Clinical-grade skin analysis in 30 seconds. Upload a selfie to analyze 14 concerns, simulate healed results, try virtual makeup, and get AI-curated product routines from an AI dermatologist.

Comment

Inspiration

Dermatology consultations are expensive and inaccessible to many, while e-commerce is flooded with generic advice. We wanted to build a bridge: a privacy-first tool that provides clinical-grade skin analysis and connects users to the right products instantly. Inspired by the precision of Perfect Corp's AI, we set out to create more than just a scanner—we built a full-cycle digital consultant.

What it does

SkinIQ transforms a simple selfie into a comprehensive skin health report. It analyzes 14 specific concerns (from wrinkles to acne), generates a "healed" simulation to show potential results, and even allows a virtual makeup try-on. Beyond analysis, it uses Gemini 2.0 to curate a personalized AM/PM routine with real products from Open Beauty Facts and features a contextual AI Consultant to answer follow-up skincare questions.

How we built it

The project is built with a modern, high-performance stack:

Frontend: React and Vite for a lightning-fast UI, utilizing a custom Canvas pipeline to auto-upscale images for maximum API accuracy. Intelligence: Orchestration of 5 distinct APIs—Perfect Corp (Analysis, Simulation, VTO), Google Gemini 2.0 (Logic & Chat), and Open Beauty Facts (Commerce). Privacy: A "Local-First" architecture where image history and scores stay in the user's localStorage, ensuring zero server-side storage of biometric data.

Challenges we ran into

One major challenge was orchestrating the high-latency analysis and recommendation steps without losing the user. We solved this by implementing a multi-stage loading state that provides real-time feedback. Additionally, building a zero-dependency client-side Markdown renderer for the AI Chat ensured the app remained lightweight and portable.

Accomplishments that we're proud of

We are particularly proud of the AI Consultant integration. It doesn't just give generic answers; it "knows" your skin scores and routine context, providing truly personalized advice. We also successfully implemented a client-side PNG report generator that creates beautiful, shareable cards using only native Browser APIs.

What we learned

We learned the critical importance of image preprocessing for AI—our auto-upscaling pipeline significantly improved the consistency of the Perfect Corp HD Analysis results. We also deepened our expertise in prompt engineering to ensure Gemini's recommendations were both clinical in tone and safe for users.

What's next for SkinIQ

We plan to integrate Real-time Video Analysis for even more granular tracking and a Product Comparison Engine that can scan a user's existing shelf and suggest better alternatives based on their evolving SkinIQ score.

Built With

Share this project:

Updates