USEFUL
The evidence supports a useful assessment under the applicable rules.
SEE WHAT'S
INSIDE.
Scan the label. Read the evidence. Get a direct verdict.
ENTER THE X-RAY↘INGREDIENT X-RAY SYSTEM
THE FRONT SELLS.
THE BACK TELLS.
Ingredence reads the part designed to be ignored.
FIVE ANSWERS.
ZERO BLUR.
Each verdict is written directly and backed by the evidence used to reach it.
NEUTRAL
The evidence supports neither a useful nor a harmful verdict.
HARMFUL
The evidence confirms harm under an applicable rule.
DANGEROUS
The evidence confirms a direct danger and identifies the reason.
NOT ENOUGH DATA
The label or evidence is incomplete. Ingredence asks for what is missing.
AI READS.
RULES DECIDE.
AI extracts and structures label facts. A deterministic, versioned rules engine assigns the final verdict.
UNCERTAINTY STAYS EXPLICIT
EXTRACT
OCR / LABELVERIFY
CONFLICT CHECKLABEL MATCHCONFIRMED
SOURCE TRACETRACEABLE
MISSING DATAEXPLICIT
VERDICT
RULE ENGINERULE VERSIONFOOD / 1.0
The final result is assigned by a deterministic rule, not by the AI model.
THE VERDICT
IS NOT
FOR SALE.
NO ADS
NO ALTERNATIVES
NO PAID INFLUENCE
Confirmed harm is called harmful. If evidence is incomplete, the answer is not enough data.
THE LABEL ENDS HERE.
UNDERSTANDING STARTS NOW.
EXPLORE THE VERDICTS↑
NATIVE FORiPHONEANDROIDHARMONYOS NEXT