01
Choose the health goal
Autumn starts from the condition you selected instead of assuming one universal idea of healthy.
Methodology
How Autumn builds a grocery score.
Autumn starts with your condition, reads the product data it can verify, and lets safety rules interrupt a recommendation when they should.
Four checks
The score moves through four gates.
Condition comes first. Then Autumn reads the product, checks for risk, and only then ranks the next swap.
02
Read the product truth
The score uses the nutritional and ingredient data Autumn can actually verify on that retailer.
03
Apply safety gates
Medication, allergen, and condition rules can block or interrupt a recommendation before swaps appear.
04
Rank the next move
Autumn only surfaces a swap when the improvement is strong enough to be worth the attention.
Signals and safety
Warnings can outrank a good nutrition score.
Autumn weights signals by condition, but keeps the result readable: show the score, flag the risk, and stay honest when the data is thin.
SugarSaturated fatSalt and sodiumFibreProteinEnergy densityPotassium where coverage allows
Hard blocks
Clearly unsafe pairings are excluded instead of scoring well by accident.
Strong warnings
If a product stays visible, the warning still leads the action lane.
Advisories
Informational relationships stay visible without pretending to be stronger than they are.
Source families
NICE and BNF
Condition and medication framing for warnings, weighting, and safety decisions.
Retailer and label data
The live nutritional and ingredient information visible on supported grocery sites.
Food composition references
Fallback and enrichment support where retailer coverage is incomplete.
Autumn safety rules
Explicit hard blocks, warnings, and advisories layered on top of the score itself.
More trust details
Want the privacy and coverage details too?
The trust page covers data privacy, retailer coverage, and what Autumn does not claim.