A food photo can never guarantee perfect nutrition values. Real food varies by portion size, ripeness, preparation, recipe, brand, and what is visible in the image. Crmbls AI is built to make nutrition tracking more transparent and reviewable - not to pretend every scan is a lab result.
Crmbls AI uses AI to identify visible foods, ingredients, meal structure, and likely portions from your photo.
When suitable matches are available, nutrition values are calculated by Crmbls AI using BLS 4.0 nutrient data and estimated or confirmed gram amounts.
Before a result becomes part of your log, you can review and adjust food names, ingredients, and gram amounts in supported flows.
Nutrition values are not fixed for every real-world food. A banana can vary by ripeness. Cooked food can differ from raw food. Recipes, oils, sauces, water loss, portion size, and preparation method can all change the final values.
That is why Crmbls AI uses accuracy and source labels as transparency signals - not as a promise that every value is exact.
BLS 4.0 / Bundeslebensmittelschlüssel is a structured nutrient data source published by the Max Rubner-Institut. Crmbls AI uses BLS-based data as one foundation for nutrition calculations, then maps and calculates app values from that data and the estimated or confirmed gram amount.
Crmbls AI calculates the app values. The Max Rubner-Institut does not endorse, support, sponsor, or have an affiliation with Crmbls AI.
Scan results can still vary because photos, portions, food matching, recipes, and preparation methods can vary. Crmbls AI is built to make that process easier to review, not to hide uncertainty.
