Our Scoring Methodology
Every Checkit score is powered by AI analysis of real ingredient data, cross-referenced against international food safety databases.
The Checkit Score (0-100)
- 80-100 (Pass) — Clean ingredients, minimal harmful additives
- 40-79 (Warning) — Some concerning ingredients
- 0-39 (Fail) — Multiple harmful ingredients detected
Scoring Factor Breakdown
- Harmful Additives — 30%
- Seed Oils — 15%
- Sugar/HFCS — 15%
- Processing Level — 15%
- Nutritional Balance — 15%
- Allergen Transparency — 10%
AI Analysis Pipeline
- Ingredient extraction and normalization
- Database cross-reference (FDA, EFSA, Health Canada)
- AI safety scoring via GPT-4o-mini
- Weighted score computation
- Quality validation and anomaly detection
Data Sources
- OpenFoodFacts — 800K+ US products
- USDA FoodData Central — 380K+ products
- FDA Recall Database — Real-time monitoring
- International regulators — EU EFSA, Health Canada, FSANZ, FSSAI
FAQs
How is the Checkit health score calculated?
The score (0-100) analyzes ingredients against six weighted factors: harmful additives (30%), seed oils (15%), sugar/HFCS (15%), processing level (15%), nutritional balance (15%), and allergen transparency (10%).
What data sources does Checkit use?
We use OpenFoodFacts, USDA FoodData Central, FDA recall databases, and international regulatory authorities including EU EFSA and Health Canada.
How often is scoring data updated?
Our database updates continuously through automated pipelines. FDA recalls are monitored in real-time.
Can I trust the score for allergy decisions?
The score flags all Big 9 allergens and cross-contamination warnings, but always read the full label for severe allergies.