Health Intelligence
How Enbodie Sources Evidence and Grades Products
Darren Crowder · · 5 min read
A grade is only worth trusting if you can see how it was reached. Enbodie gives each product one score from A to E, built from six pillars. These are the rules we work by when we collect the evidence behind it.
Last reviewed 10 October 2026.
1. We go to the original source
We draw on regulators (such as the SCCS, ECHA, the MHRA, EFSA and the US FDA), government bodies (such as the US National Library of Medicine) and peer-reviewed science. When we cite a limit, a ban or a study, we use the document the regulator or researcher actually published. We cite these bodies as sources. We are not affiliated with them, and they have not endorsed Enbodie.
We do not use crowd-sourced scores, forums, reviews, blogs or AI-written summaries of research, even to double-check. Preprints, retracted papers, and animal or lab-dish studies are not used to make a claim about you. The one thing we take from the maker is what is printed on its own label, such as the ingredient list.
2. Every limit belongs to the place that set it
Rules about ingredients are not the same everywhere. California differs from New York, Scotland from England, and the European Union from the United Kingdom.
So every limit we hold is stored with the regulator that issued it and the exact place it applies to, at the level that issued it. A state’s rule stays with that state and is not stretched to cover the whole country. A rule set for a bloc such as the European Union stays with the bloc and is not copied onto each member country.
A product is graded only against law that reaches where it is sold. We do not borrow a number from another country to fill a gap, because a borrowed limit looks authoritative but is not the law where you live. Today the app shows a grade made under your own country’s law for the UK, the United States and Canada. For other countries it shows a grade from a market we do hold, and says in plain words whose law that grade was made under and that we hold no rules for your country yet. The one exception to “your law only” is a ban, covered below.
3. The law comes first
Where a regulator has already stated the answer, such as a maximum concentration, a permitted form or an outright ban, that is the answer. We do not re-work a number the law has already fixed, and we do not give a softer verdict than the law does.
We also label whose line it is. Where we go beyond the law, such as an Enbodie estimate, or a combined-use figure that no regulator publishes, the app says it is Enbodie’s own reasoning and not a regulator’s ruling.
4. A ban anywhere is a ban everywhere
If any regulator in the world has banned an ingredient for a kind of product, we flag it on every product of that kind that contains it, wherever it is sold and wherever you live. A cosmetics ban is applied to cosmetics, and a supplement ban to supplements. Your location decides which regulator we name. It does not decide whether you are told.
We tell you so that you can make an informed decision. Where a ban is not a ruling by your own regulator, we say so plainly, as Enbodie’s recommendation. What you do with it stays your choice.
5. Every serious warning has a paper trail
Every red or critical verdict has to trace back to one of three things: a named regulator’s published document; an Enbodie safety threshold with its reasoning written down and labelled as ours; or a systematic review of peer-reviewed studies. A single study is not enough, however good. A check runs every night and flags any verdict that is missing its trail.
6. Safety limits come first
Safety limits come first, not what sells best or who is advertising or sponsoring. The measure depends on the product. For cosmetics it is the Margin of Safety: the dose with no observed effect, divided by what you are exposed to. A margin below 100 is below the threshold regulators use. Supplements are checked against upper intake limits, foods against acceptable daily intakes, and medicines against their licence.
7. Estimates and gaps are said plainly
Where nobody has measured a safety value, we use cautious methods regulators accept, such as reading across from a closely related chemical or a default set by chemical class, and we label the result as an estimate. In the app, each ingredient shows whether its figure is measured, reviewed or estimated. Where even that is not possible, we show that there is not enough data rather than a number. We do not dress up an estimate as a measurement.
Our sourcing and environmental pillars are newer, and the data behind them is still being built.
8. AI does not decide any grade
Every verdict is worked out from the data by fixed rules before AI sees it. The sums are fixed arithmetic, not an AI’s opinion. AI only turns the finished result into plain language.
What this does not mean
Enbodie does not diagnose medical conditions and is not a substitute for professional medical advice. A grade is a way to understand a product, and a reason to ask a question of a pharmacist, doctor or optician when something matters to you.
If you think we have got something wrong, write to [email protected]. A correction that points to a primary source is the most useful thing you can send.