Short version: the score describes label transparency, evidence-to-claim context and value disclosure. It is not a clinical rating, safety clearance, laboratory test or prediction that a product will work for you.
The 100-point label score
| Component | Maximum | What earns points |
|---|---|---|
| Disclosure | 48 | A usable ingredient list, multiple named ingredients, detectable amounts and uses/servings. Proprietary blends lose disclosure points. |
| Evidence context | 38 | Ingredients found in our cited library, especially when their researched outcomes overlap the promises selected by the user. Grades add different, modest weights. |
| Value disclosure | 14 | A price and number of uses that let the browser calculate price per use and estimated 30-day cost. |
How verdicts map to scores
- 75–100 — Receipts provided: relatively detailed information and relevant library matches.
- 55–74 — Some substance: useful detail with material gaps or uneven evidence-to-claim fit.
- 35–54 — Mixed signals: a partial picture that deserves a slower look.
- 0–34 — More vibe than detail: too little disclosed information for a useful comparison.
What the evidence matching does
The browser looks for exact ingredient names and aliases in pasted label text. It then checks whether the outcomes attached to those ingredients overlap the promises you selected. The library is deliberately small and source-linked. A non-match means “not covered here”, not “no evidence exists”.
Evidence labels such as established role, promising, limited and context matters describe our summary of the cited source. An ingredient’s research cannot prove a multi-ingredient finished product has the same effect.
What the score cannot see
It cannot verify identity, purity, contamination, manufacturing quality, dose accuracy, bioavailability, whether a study used the same formulation, or whether a product is appropriate for an individual. It does not account for diagnoses, medicines, pregnancy or personal risk.
Barcode records
Barcode digits are decoded on the device and sent directly to the community-maintained Open Facts ecosystem. Those records can be incomplete or wrong. They are used to prefill the form, never as independent evidence, and the physical pack should be treated as the source of truth.
Commercial independence
The scoring module receives product-label fields only. It has no input for sponsors, affiliate commission, advertising or brand relationships. Commercial placements are configured separately and must be visibly labelled. Read our commercial policy.
Version and corrections
This is methodology version 0.1, last materially reviewed 26 August 2026. The library and rules should be reviewed as evidence and the product evolve. A public corrections/contact address is a required launch task before production.