In short
A research association, not a verdict. The score describes how a food sits in published nutrition research on inflammation. It says nothing certain about what that food will do in your body.
Built from published data. The numbers come from a peer-reviewed dietary index and public food-composition databases, not from opinion and not from AI.
The same answer every time. The same food always gets the same score under the same scoring version. No AI model ever chooses a number.
What the number means
The score runs from 0 to 100. A higher number means a higher inflammatory association in the research the score is built on; a lower number means a lower one.
It is a comparison, not a measurement. Chronly scores about a hundred food categories, and each one's number shows where it sits relative to the others. It is not a quantity of anything in your blood or your body.
To make it easier to read, every score falls into one of three bands:
These words describe where a food sits on our scale. They do not mean a food is good or bad, or safe or unsafe for you.
Where the number comes from
The nutrients in a food. The base of every score is the Dietary Inflammatory Index, a published research tool that gives 45 food components a weight according to how they related to inflammatory markers across nearly two thousand studies. Fibre and many plant compounds carry weights in one direction; saturated fat carries a weight in the other. We apply those published weights to each food category's nutrient profile, measured per 100 calories, using the United States Department of Agriculture's food-composition data. Components with no reliable public data are left out rather than guessed.
How processed a food is. The index cannot see some things the research has linked with inflammation: ultra-processing, processed meat and added sugar. Foods in those groups carry an additional, fixed weight, reviewed before it was used.
For a packaged food (barcode). We read the product's ingredient list from Open Food Facts, an open database of food products. For US products it does not hold, we use the USDA's branded-food records. Each ingredient is matched to one of our food categories, and the product's score is the average of its ingredients, weighted by how many calories each contributes. Its processing level, using the NOVA food classification, and any food additives on the label are also counted; additives are weighted by how strong the evidence about them is. If we cannot recognise at least 90% of the label, we do not give it a full score rather than guess.
For a photo of a meal. An AI image model looks at the photo and names what is on the plate, choosing only from our fixed list of food categories, and says whether each item is a main part of the dish, a side or a small amount. That is all it does. It never produces, adjusts or explains the score. The score is then calculated the same way as for any other food, weighting main items most and small amounts least. A photo cannot show additives, so a photo scan is less complete than a barcode scan of the same food.
Checked before it reaches you
The judgement calls behind the score — the processing and additive weights, how each ingredient on a label is matched to a food category, the recipes used for common dishes, and the few foods whose number we set by hand where the data could not be right — and every sentence the app shows to explain a score were reviewed and signed off by clinicians before they were used.
Each score is stored with the version of the method that produced it. When the method improves, the version changes. A scan you have already saved keeps the score it was given, and we do not change past results without telling you.
What the score cannot tell you
It describes a food, not you. It is based on research about groups of people. How a food affects you depends on things the score cannot see, such as the amount you eat, how it was prepared, your medication and your own body.
It is the same for every condition today. Chronly does not yet adjust the score for your diagnosis. The research differs between conditions, and we would rather show one honest number than a tailored one we cannot yet support.
It does not know portion sizes. Chronly does not estimate grams from a photo, because that estimate is too unreliable to build on.
The evidence on additives is still limited. Much of it comes from laboratory and animal studies, and the few studies in people at the amounts found in food have not shown an effect on inflammation. We include additives because of the wider research, and the app says what kind of evidence sits behind each one.
Some foods cannot be scored. When too little of a meal or a product can be identified, Chronly says so, or shows a provisional result, rather than inventing a number.
Your patterns are a separate thing
The food score is about food in general. Chronly's Insights are about you: they compare what you have logged over time, such as sleep, stress, meals and symptoms. They only appear after several weeks of logs, they describe what you have logged and when, and they never claim that one thing caused another.
Sources
Chronly does not replace professional medical advice.
It helps you notice patterns in what you log. It does not diagnose, treat or prevent any condition. Talk to your doctor or care team before changing your diet or your treatment.
Questions about this page: contact@chronly.app