Computational Formulation Intelligence
One computational model holds together chemistry, the biology of skin, hair and mouth, formulation craft and the scientific literature. Every prediction arrives with its proof and its limits — always with the reasoning.
Constraints from Reg. (EC) 1223/2009 in the registry · every value with its source · verifiable predictions, not certifications
Predictive cosmetic science
The destination of use is not a label: it decides which endpoints make sense, how doses are read and which warnings matter. Each area says where it stands.
full model
The core of the model: 32 skin endpoints across efficacy and safety, each with its evidence.
Analyze for this areaskin model + 6 dedicated
Body skin: the 32 Skincare endpoints plus six of its own — sweat, odor, cellulite, stretch marks, localized fat, unwanted hair.
Analyze for this areadedicated model active
24 endpoints, shared and hair-specific: growth, shedding, fiber, dandruff, color.
Analyze for this areadedicated model active
A mouth has rules of its own: 16 endpoints, from plaque to the microbiome. The first directions are already in the engine.
Analyze for this areafull model + 6 solar
Skin under the sun: the full model plus six dedicated endpoints — UVB and UVA protection, photostability, UV damage.
Analyze for this areaessential model active
The intimate mucosa: pH and microbiome, with the shared tolerance endpoints. Dedicated literature is scarce, and we say so.
Analyze for this areaessential model active
The nail plate: brittleness and growth, plus the skin around it.
Analyze for this areaDocumentary base and coverage
What sits under every prediction: how many ingredients are recognized, for how many we know the direction of effect, how much literature was read recently.
A growing base: new literature comes in every 30 minutes and is read every hour.
Fenoteca — the open library of ingredients
Every entry separates the documented from the not-yet-documented: a missing datum stays missing — it lowers confidence, it never becomes a zero effect. Above: the eight ingredients with the most evidence in the registry.
The three-layer architecture
The Knowledge Base gathers, orders and updates the scientific data. What we build on top of it is the heart of the project.
INCI, literature, safety, dose-response, interactions, physico-chemical properties. It updates itself, every day.
The complete formula — doses, pH, vehicle, use, skin profile — becomes a prediction of efficacy, compatibility and tolerability, with the reasoning.
The digital model of an individual skin, built over time: evolution tracked and — once validated — simulation of future scenarios.
Describe the formula in your own words. See which INCI names are recognized, where an effect was measured in a study and where it was not, how pH and vehicle change the reading, which known interactions touch your formula — with confidence reported separately from the value: every number traces back to its coefficient and its evidence.
Analyze a formula«Missing evidence is not an effect equal to zero.»
No blacklists, no alarmism: we assess ingredients and formulas in their context — dose, vehicle, evidence and declared limits. It is the first of the seven rules every report obeys.
The seven rules of every report