Order: (1) weighted sum of the category’s field scores, (2) concern multiplier (cosmetic leave-on), (3) xylitol-in-dog cap to 0, (4) category base cap, (5) conditional caps, (6) clamp to 0–100. Two caps that both apply: the lowest wins.
Category field weights
Each category’s weights sum to 1.0. The AI fills the field scores; this table is the arithmetic.
food
| Field | Weight |
|---|
| additive safety |
25% |
| sugar concern |
20% |
| ingredient quality |
20% |
| nutritional value |
15% |
| allergen clarity |
10% |
| preservative load |
10% |
cosmetic
| Field | Weight |
|---|
| endocrine disruptors |
30% |
| carcinogen irritants |
25% |
| fragrance transparency |
15% |
| surfactant harshness |
15% |
| active ingredients bonus |
15% |
supplement
| Field | Weight |
|---|
| drug interaction risk |
30% |
| megadose toxicity |
25% |
| bioavailability quality |
20% |
| filler quality |
15% |
| label transparency |
10% |
personal-care
| Field | Weight |
|---|
| absorption zone risk |
30% |
| toxic chemical presence |
25% |
| ingestion risk |
20% |
| fragrance load |
15% |
| microbiome impact |
10% |
household
| Field | Weight |
|---|
| acute hazard level |
30% |
| mixing danger |
25% |
| voc respiratory risk |
20% |
| child pet safety |
15% |
| env persistence |
10% |
baby
| Field | Weight |
|---|
| heavy metal risk |
30% |
| developmental toxins |
25% |
| age appropriateness |
20% |
| additive sensitivity |
15% |
| allergen safety |
10% |
pet
| Field | Weight |
|---|
| pet toxin detection |
35% |
| protein quality |
25% |
| preservative safety |
20% |
| species life stage fit |
20% |
tobacco
| Field | Weight |
|---|
| additive toxin load |
40% |
| delivery mechanism risk |
35% |
| relative harm context |
25% |
unknown
| Field | Weight |
|---|
| visible hazard signals |
40% |
| universal red flags |
35% |
| label readability |
25% |
Caps
- household cannot score above 55.
- tobacco cannot score above 15.
- critical drug interaction caps the overall score at 30.
- infant honey caps the overall score at 20.
- baby age inappropriate caps the overall score at 40.
- xylitol dog caps the overall score at 0.
Cosmetic leave-on products multiply the concern (100 − score): leave-on 1.4, lip 1.3, leave-on hair 1.2, rinse-off 1.0.
What this is not
A score is not a lab assay, a medical opinion, or a claim that a product is “safe” or “unsafe.” It is a deterministic combination of public regulatory signals the model extracted from the photo. Personalised Pro scans reuse the same arithmetic after the model applies your profile to the field scores.
How the analysis pipeline works →