How we read your skin.

478 facial landmarks · 7 skin zones · 1,200+ ingredient profiles. Built on peer-reviewed dermatology AI research.

What backs this up

• AI vs. dermatologist diagnostic accuracy

  Systematic review and meta-analysis. AI sensitivity 86.3%, specificity 78.4% — comparable to specialists.

  npj Digital Medicine · 2024 ↗
• Facial severity scoring across regions

  AI-graded scoring strongly correlates with dermatologists across forehead, periorbital, nasolabial, lower-face ptosis, and diffuse redness.

  MDPI Cosmetics · 2024 ↗
• Skin-hydration AI accuracy

  Capacitive-imaging AI achieves 83.8% accuracy for skin hydration assessment.

  MDPI literature review ↗
• AI dermatology general performance

  Recent review reports AI dermatology accuracy reaching up to 99% on several condition classes.

  PMC · 2025 ↗

Fairness disclosure

Dermatology AI models — including the underlying methods we rely on — perform stronger on lighter Fitzpatrick I–III skin (AUROC ~0.89) than on darker IV–VI (AUROC ~0.82). We document this openly and are actively working toward more diverse training data.