Skin Moment

Inspiration

You don't wonder about your skin in the abstract. You wonder right before you buy something, right after a breakout, right in front of the mirror. Most skincare apps treat skin analysis as a novelty — a score you look at once and forget. We wanted to build something that meets you at the actual decision moment: should I buy this?

What it does

SkinMoment is a three-part experience built on YouCam Skin AI:

1. Skin Snapshot — Upload a selfie and get 8 dermatologist-grade metrics (moisture, oiliness, redness, acne, spots, wrinkles, texture, dark circles) via the YouCam Skin AI API.
2. Right-Now Plan — A Groq-powered coach turns your two lowest-scoring metrics into 3 concrete, non-medical actions you can take today.
3. Before-You-Buy Check — Paste any product's name and description. SkinMoment weighs it against your current snapshot — not a star rating — and returns a verdict: good fit, caution, or poor fit, with reasoning tied to your actual scores.

How we built it

• Next.js 14 (App Router) + TypeScript + Tailwind for the frontend
• Full YouCam V2 flow: file upload → pre-signed URL → skin-analysis task → polling → normalized 8-metric result
• Groq (Llama 3.3 70B) reasoning layer that turns raw scores into plain-language guidance and a purchase-fit verdict
• Demo mode with realistic mock scores so the full pipeline is testable end-to-end even without live credentials

Challenges we ran into

Getting the auth flow right. Perfect Corp's V1 docs describe an id_token/client_id exchange that assumes credentials we didn't have; we switched to the V2 API, which just uses the API key directly as a Bearer token — much simpler once we found it.

What we learned

That the real product opportunity isn't the skin score itself — it's connecting that score to a decision the person is already trying to make (a purchase).

What's next

Webhook-based analysis instead of polling, saving snapshot history to track trends over time, and extending the purchase-check to parse ingredient lists automatically from a product photo.

Built With

• groq
• llama-3.3
• next.js
• tailwind-css
• typescript
• youcam-api