Forma Research

Skin Intelligence,
Explained.

The science behind every scan — from how Forma reads your complexion to why shade matching is more than picking a color.

Our Principles

Observable, not assumed

Forma reads what the camera can actually see. Skin type labels like 'oily' or 'dry' are replaced with measured signals: surface variance, specular reflection, and regional hue.

Local by design

All vision processing happens in your browser. Your camera frames never leave your device. The AI consultation uses only anonymized signals, not images.

Calibrated recommendations

Every formula recommendation is traceable back to your specific scan measurements. We show you the reasoning, not just the result.

Deep Reads

Foundations4 min read

Complexion Science

The Undertone Spectrum

Warm, cool, and neutral undertones arise from the ratio of oxyhemoglobin to melanin in the skin. Forma maps this by analyzing hue deviation in the red-green channel across three facial regions — and it matters more than your surface shade.

Color5 min read

Methodology

How Shade Matching Works

Matching foundation shade is a two-variable problem: depth (light to deep) and undertone (warm, cool, neutral). Forma reads both simultaneously using calibrated luminance sampling and cross-references against 200+ formulas to find your closest fit.

Formulation6 min read

Skin Intelligence

Texture, Hydration & Finish Direction

Surface texture signals hydration levels, pore density, and sebum activity — all of which determine whether a dewy, satin, or matte finish will perform best on your skin. Forma reads surface variance in micro-regions to map this in real time.

Application4 min read

Personalization

Coverage That Fits

Light, medium, and full coverage serve different visibility goals and skin conditions. Forma factors in your visible concern profile — redness, hyperpigmentation, texture — to recommend coverage that achieves an even result without masking.

AI7 min read

Technology

Inside Forma's AI Model

Forma runs a local computer vision pipeline in your browser — no server upload, no stored images. Face landmark detection anchors region-specific sampling, and a multi-layer analysis model combines local inference with AI consultation to produce your results.

Skincare5 min read

Skin Literacy

Understanding Your Skin Concern Profile

Redness, hyperpigmentation, dehydration, fine texture, and large pores each present differently under varied lighting. Forma identifies and ranks your visible concern profile so your routine addresses what's actually present — not a general category.

See the science in action.

A 30-second scan — Forma reads your complexion and shows you exactly what it found.

Begin Your Scan