Who scores the products on MiauMiau?

Every product is scored by the MiauMiau Nutritional AI against the rubric on this page. We are not veterinarians — we are a transparent scoring system that compares the published ingredient and nutritional data of each product against AAFCO, FEDIAF, and NRC nutrient profiles. Our methodology is open-source above; the model versioning is logged in the version history block.

What does the 0-5 score mean?

1.0 to 1.9 is poor (significant filler load or unsafe additives). 2.0 to 2.9 is below average (typical mass-market dry kibble). 3.0 to 3.9 is good (above-average ingredient quality, named meats). 4.0 to 4.9 is very good (whole foods, conservative additive use, balanced nutrition). 5.0 is reserved for the rare product that excels on all four dimensions simultaneously.

How is the composite score calculated?

We compute four sub-scores (each 1.0-5.0) for ingredient quality, protein adequacy, filler load, and additive safety. The weighted average is taken (35%/30%/20%/15%), then capped at lowest_subscore + 1.5 — this prevents one weak dimension from being washed out by three strong ones. The final number is floor-rounded to one decimal and stored as the canonical quality score.

Are scores ever influenced by money?

No. Scores are mechanical: rubric in, score out. We may in the future accept paid placement for products that already score above a public floor, but the score itself is never for sale. If that policy ever changes, the change will be noted on this page in the version history.

How often are scores updated?

Continuously. Every newly ingested product is scored within hours of arrival. Existing scores are re-run whenever the rubric changes (see version history) or when we ingest a fresh ingredient panel from a manufacturer's site. Each product page shows the exact date its current score was computed.

Methodologyv2.1 · Updated April 29, 2026

How every score is calculated

Every product in the MiauMiau catalog is scored on a 0–5 scale using a four-dimension rubric grounded in AAFCO, FEDIAF, and NRC feline nutrition standards. This page is the single source of truth — when scores change, it is because the rubric below changed, and the change is logged in the version history at the bottom of the page.

The four dimensions

Each sub-score is computed independently, then combined into the final composite. Weights sum to 100%.

Ingredient quality

35%

Whole named meats (chicken, salmon, duck) score highest. Generic 'meat meal' or 'animal by-product' lose points. The first three ingredients matter most because pet-food labels list ingredients by pre-cooking weight.

Protein adequacy

30%

Cats are obligate carnivores. We score the dry-matter-basis protein percentage against AAFCO 2024 maintenance minimums (26% adult / 30% growth) and benchmark against the per-category median in our database.

Filler load

20%

Corn, wheat, rice, soy, pea protein concentrates and starches contribute calories without species-appropriate nutrition. Higher filler share = lower score, weighted heavier in dry food (where fillers are most common).

Additive safety

15%

Artificial colors (Red 40, Yellow 5/6), BHA/BHT preservatives, propylene glycol, and rendered fat sources are penalized. Mixed tocopherols, rosemary extract, and named natural preservatives are neutral.

The formula

The composite score is a weighted average, capped to prevent a single mediocre dimension being washed out by three strong ones.

weighted = 0.35 * ingredient_quality
         + 0.30 * protein_adequacy
         + 0.20 * filler_score
         + 0.15 * additive_safety

composite = min(weighted,
                lowest_subscore + 1.5)

final = floor(composite * 10) / 10   // 0–5 scale, 1 decimal

The cap is load-bearing: without it, a product scoring 5/5 on three dimensions and 1/5 on additive safety would still score ~4.4. With it, that same product caps at 2.5 — which is what most cat owners would intuitively expect from a label that contains BHT or artificial color.

Primary sources

The standards we benchmark against. None of these are paid relationships — we cite them because they are the authoritative published references for feline nutrition.

AAFCO 2024 Cat Food Nutrient Profiles
The Association of American Feed Control Officials publishes minimum nutrient requirements for cat food. Used as the floor for our protein-adequacy sub-score.
FEDIAF Nutritional Guidelines for Complete and Complementary Pet Food
European federation of pet food industry's published nutrient profiles. Used as the EU-equivalent reference where it differs from AAFCO (e.g., higher minimum protein recommendations).
NRC Nutrient Requirements of Dogs and Cats (2006)
National Research Council's primary academic reference for feline nutrition. We cite it for the few requirements where AAFCO and FEDIAF disagree.

What this is — and isn't

MiauMiau is not a veterinary clinic. We do not employ staff veterinarians. The scores on this site are produced by an AI scoring system — the MiauMiau Nutritional AI — applying the rubric on this page to the published ingredient and nutritional data on each product.

We are a published, dated, version-controlled rubric. You can read this page, you can read every score we've ever produced, and you can audit the model version that produced it (visible on every product detail page). The same input always produces the same output, and the rubric changes are tracked below.

Talk to a vet for medical advice. If your cat has a medical condition, allergy, or special dietary need, no online scoring system — including this one — is a substitute for a real DVM examining your cat. Our scoring is for comparing products, not for diagnosing.

Version history

Every public rubric change. When a number on this page moves, the change is added here with a date.

v2.1April 29, 2026
- Tightened composite cap from min_subscore + 2.0 → +1.5 to prevent a single mediocre dimension being washed out by three high ones.
- Migrated scoring model to OpenRouter for routing flexibility — same rubric, model is now per-workload editable.
v2.0April 18, 2026
- Reweighted from 0.40/0.25/0.20/0.15 → 0.35/0.30/0.20/0.15 to give protein adequacy more weight after audit feedback.
- Added Phase 2 cross-region edition support — single canonical product can now be scored from translated ingredient panels in any market.
v1.0March 14, 2026
- Initial public scoring rubric: 4 sub-scores, weighted average, 0-5 scale.

60 SECONDS · FREE

See the rubric in action

Browse the full catalog, every product graded against the rubric on this page.

Browse the food database

AI-SCORED · EU-NATIVE