The Dtoxify Score

The Dtoxify Score is a composite rating from 0 to 100 that summarizes the overall health profile of a food product. It is designed to give consumers a quick, intuitive measure of product quality that combines multiple dimensions of food science into a single number.

The score integrates three established, evidence-based assessment frameworks:

Nutri-Score

A–E

Nutritional quality based on EU methodology

NOVA Group

1–4

Processing level classification

Additives

0–n

Number & severity of flagged chemicals

Nutri-Score Component

Nutri-Score is a front-of-pack nutrition label developed by the French National Research Agency (INRA) and adopted in France, Belgium, Germany, Spain, the Netherlands, Luxembourg, and Switzerland. It rates products from A (best) to E (worst) by calculating a composite score based on negative nutrients (energy, sugars, saturated fatty acids, sodium) and positive nutrients (protein, fiber, percentage of fruits/vegetables/nuts/legumes). When available directly from our databases, we use the pre-calculated Nutri-Score grade. When not available, we note it as "Unknown."

NOVA Classification Component

NOVA is a food classification system developed by researchers at the University of São Paulo, Brazil. It classifies food products into four groups based on the extent and purpose of food processing: Group 1 (unprocessed or minimally processed foods), Group 2 (processed culinary ingredients), Group 3 (processed foods), and Group 4 (ultra-processed food and drink products). Strong epidemiological evidence links ultra-processed food consumption (NOVA 4) with increased risk of obesity, cardiovascular disease, type 2 diabetes, and other health conditions.

Additive Safety Component

Dtoxify identifies and flags food additives based on safety assessments from regulatory bodies. Additives are cross-referenced against evaluations by JECFA (Joint FAO/WHO Expert Committee on Food Additives) which has evaluated 6,552 substances with established Acceptable Daily Intake (ADI) values, IARC (International Agency for Research on Cancer) classifications for carcinogenicity, and EFSA (European Food Safety Authority) scientific opinions. Flagged additives are those that have been classified as potentially carcinogenic by IARC (Group 2A or 2B), are restricted or banned in certain jurisdictions, or have known adverse effects documented in peer-reviewed literature. Each flagged additive is shown with a risk level (low, moderate, high) and a brief plain-language explanation of the concern.

Transparency principle: Every flagged chemical includes the specific reason for concern and the regulatory body that raised it. We never flag ingredients without citing the scientific basis.

Multi-Database Product Lookup

Unlike single-source apps, Dtoxify queries multiple authoritative databases in a waterfall sequence to maximize the chance of finding any product worldwide. When you scan a barcode, here is what happens:

1 Open Food Facts — 4M+ products, 150+ countries
↓ not found
2 Open Beauty Facts — 62K+ cosmetics & personal care
↓ not found
3 USDA FoodData Central — 380K+ US foods
↓ not found
4 FatSecret — 1.9M foods across 56 countries
↓ not found
5 OCR Fallback — photograph the ingredients label for AI analysis

After a product is found, it passes through enrichment layers that add nutritional data from regional food composition databases — the Indian Nutrient Databank (ICMR-NIN) for Indian products and Health Canada's Canadian Nutrient File (CNF) for Canadian products. Finally, a safety layer checks the product's brand against OpenFDA's food enforcement database for active recalls.

4.4M+
Searchable products
150+
Countries covered
7
Databases integrated
6,552
Chemicals evaluated

Data Sources — Product Databases

Open Food Facts

4M+ products · 150+ countries · ODbL license · openfoodfacts.org

The world's largest open database of food products. Provides product name, brand, ingredients, nutritional data, Nutri-Score, NOVA classification, allergens, and categories. Data is crowdsourced and moderated, with AI-assisted quality checks (Robotoff). Dtoxify queries this as the primary database for every barcode scan.

Open Beauty Facts

62K+ products · Global cosmetics & personal care · ODbL license · openbeautyfacts.org

An open database of cosmetics and personal care products, maintained by the same community as Open Food Facts. Enables Dtoxify to scan and analyze skincare, haircare, and beauty products — not just food. Provides ingredient lists (INCI format), brand data, and additive identification for cosmetic products.

USDA FoodData Central

380K+ foods · United States · Public domain (CC0) · fdc.nal.usda.gov

The U.S. Department of Agriculture's comprehensive food composition database. Provides detailed nutritional profiles for branded products, SR Legacy foods, and survey data. Used as a fallback when a product is not found in Open Food Facts, particularly for US-market products. Data is in the public domain under CC0 license.

FatSecret Platform API

1.9M+ foods · 56 countries · 24 languages · platform.fatsecret.com

A comprehensive international food database covering 56 countries and 24 languages. Provides nutritional data, serving sizes, and food categorization. Used as an additional fallback layer in the waterfall, significantly expanding coverage for branded products in markets where Open Food Facts coverage is sparse.

Indian Nutrient Databank (ICMR-NIN)

1,212 foods & recipes · India · ICMR — National Institute of Nutrition

India's authoritative food composition database compiled by the Indian Council of Medical Research. Covers 528 raw food ingredients and 1,014 traditional Indian recipes with full macronutrient and micronutrient profiles. Used as an enrichment layer — when an Indian product is found with sparse nutritional data, INDB fills in the gaps with government-validated composition data.

Health Canada — Canadian Nutrient File (CNF)

5,690 foods · Canada · 89 nutrients per food · food-nutrition.canada.ca

Canada's official food composition database maintained by Health Canada. Contains detailed nutritional profiles for 5,690 Canadian foods including 13 key macronutrients and micronutrients (energy, protein, fat, carbohydrates, fiber, sugars, sodium, saturated fat, calcium, iron, vitamin C, cholesterol, potassium). Used as an enrichment layer alongside INDB — when a Canadian product lacks nutritional data, CNF provides government-validated composition data.

OpenFDA Food Enforcement

Safety alerts · United States · No API key required · open.fda.gov

The U.S. Food and Drug Administration's food enforcement and recall database. Checked automatically on every scan — when a product's brand matches an active or recent recall, Dtoxify displays a safety warning with the recall classification (Class I: serious health hazard, Class II: moderate risk, Class III: minor violation) and the reason for recall. This adds a real-time safety dimension beyond nutritional analysis.

Data Sources — Safety & Toxicology

WHO/JECFA Evaluations

6,552 substances · Established since 1956 · who.int/groups/joint-fao-who-expert-committee-on-food-additives

The Joint FAO/WHO Expert Committee on Food Additives is the international body responsible for evaluating the safety of food additives, contaminants, naturally occurring toxicants, and residues of veterinary drugs in food. JECFA establishes Acceptable Daily Intake (ADI) and Tolerable Daily Intake (TDI) values that form the basis of global food safety standards. Dtoxify uses JECFA evaluations to assess additive safety.

EFSA Scientific Opinions

European Food Safety Authority · 5,712 chemicals · efsa.europa.eu

EFSA provides independent scientific advice on food-related risks to the European Commission, European Parliament, and EU Member States. Their risk assessments cover food additives, pesticide residues, contaminants, and novel foods. EFSA's OpenFoodTox database contains information on over 5,712 chemicals. Dtoxify references EFSA opinions for European-specific additive assessments.

IARC Monographs on Carcinogens

International Agency for Research on Cancer · WHO · monographs.iarc.who.int

IARC evaluates the carcinogenicity of chemicals, complex mixtures, and other agents. Their classification system (Group 1: carcinogenic, Group 2A: probably carcinogenic, Group 2B: possibly carcinogenic) is used globally as the reference standard for cancer risk assessment. Dtoxify flags additives that fall in IARC Groups 1, 2A, or 2B.

Data Sources — Technology

Google Cloud Vision API

Barcode detection & OCR · cloud.google.com/vision

Used for barcode detection from user-submitted photos and OCR (optical character recognition) to extract text from ingredient labels. Employs a multi-pass detection strategy: raw image scan, enhanced contrast scan, bottom-60% crop, and upscaled retry. Photos are processed in transit and not stored permanently.

GPT-4.1-mini (OpenAI)

Natural language generation · openai.com

Generates human-readable scan reports in 16 languages, explains technical terms (Nutri-Score, NOVA groups, E-numbers, ADI values) in plain language, and provides personalized context based on the user's allergy profile. All AI-generated content is grounded in data from the databases listed above — never speculative health claims.

Scoring Methodology Limitations

We believe in being transparent about what our scoring can and cannot do. The Dtoxify Score is designed to be a helpful starting point for consumers, not a definitive health assessment.

Current limitations include the following considerations. Nutri-Score was designed primarily for European markets and may not perfectly capture nutritional quality of all traditional foods from other regions. NOVA classification focuses on processing level, which is one important factor but does not capture all aspects of food quality. Additive flagging is based on known regulatory assessments; new research may change the safety profile of specific chemicals. Database coverage varies across countries and product categories, with the strongest coverage in Western Europe, the United States, and India. Product formulations change over time, and databases may not always reflect the most current version of a product.

Health Disclaimer: Dtoxify scores are for educational purposes only. They do not constitute medical advice, diagnosis, or treatment recommendations. Individual responses to food ingredients vary based on genetics, health conditions, dosage, and other factors. Always consult a qualified healthcare professional for personal health decisions, especially if you have allergies, medical conditions, or take medications.

Open Source Commitment

Dtoxify is built on open data and open science. Product data from Open Food Facts is used and contributed back under the Open Database License (ODbL). USDA data is in the public domain (CC0). Chemical safety data is sourced from publicly available regulatory databases. We believe that access to food safety information should be universal, which is why Dtoxify operates on WhatsApp — the most widely used messaging platform globally — eliminating the need for app downloads or specialized technology.

Future Data Sources (2026 Roadmap)

We are actively working to expand our data coverage. Planned integrations include:

These additions will enable multi-dimensional scoring across nutrition, safety, sustainability, and personal relevance.

Contact & Feedback

Questions about our methodology? Spotted an error in our data? We welcome feedback and corrections. Message us on WhatsApp at +91 9360 765 428 or visit dtoxify.life.