From photo to personalised colour and skincare guidance in four steps — with measurement, not guesswork, at the core.
Step 1
Capture with a quality gate
Capture a live photo or upload one. MediaPipe face detection runs in your browser and scores four checks in real time — Face Detected, Good Lighting, Sharp Image, Face Centered — and analysis only starts at a 70% quality score.
Step 2
Measured skin colour
Pixels from your cheeks, forehead, and jaw are sampled, outliers removed, lighting corrected, and your depth, undertone, and natural base classified in the Lab colour space.
Step 3
Structured profile
You get a skin profile — one of 12 depth categories, Warm/Cool/Neutral undertone, one of 6 natural bases, one of 5 skin types — each with an honest confidence score.
Step 4
Personalised guidance
An AI vision model turns the measurements into dress colours with ratings, colours to avoid, outfit ideas, grooming tips, and a morning and night skincare routine.
Why the quality gate comes first
Any photo-based analysis is only as good as the photo. Instead of analyzing whatever it receives and returning a confident-sounding wrong answer, MoonSkin blocks weak inputs up front: the live camera view continuously scores face detection, lighting, sharpness, and centering, and won't start until the combined score passes 70%. The four indicators show you exactly which check is failing, so fixing a bad photo takes seconds — turn toward a window, hold still, move your face into the oval.
Why skin tone is measured, not asked
The Lab colour space separates lightness (L*) from colour cast (a/b) — which is exactly the separation between skin tone depth and undertone. MoonSkin samples real facial skin pixels, drops highlights and shadows, applies a bounded lighting correction, and then reads depth from L* alone and undertone from a/b alone. That is why the result is one of 12 evenly spaced depth categories from Very Fair to Very Deep, rather than a vague “medium”. The AI vision model never overrides the measured tone — it describes what it sees and builds the styling guidance on top.
Everything ends in a report you can keep: dress colours with hex codes and star ratings, colours to avoid, grooming tips, a morning and night routine, and PDF export. Reports save to your browser only — see the FAQ for the full privacy picture.