Supplement marketing has one volume setting. Everything is "clinically proven," "science-backed," "research shows"; whether it rests on fifty randomised trials or a single study in twelve rats. Eclo exists partly to undo that flattening. Every compound in the app carries an evidence grade, and the grade is the honest answer to one question: how good is the evidence that this does what it is taken for?
This page explains how those grades are assigned, so you can decide how far to trust them.
The most important thing to understand is that a supplement is not graded as a whole. It is graded separately for each goal it is taken for. Magnesium has reasonable evidence for some uses and very little for others, and a single mark would hide exactly the distinction that matters. So in Eclo the same compound can read Moderate for one goal and Weak for another. A grade is always a grade for a use.
It follows that where a compound has no real evidence relationship with a goal, we leave it ungraded for that goal rather than inventing a weak one. The absence of a grade is information too.
Strong. Multiple well-conducted human trials or meta-analyses, pointing the same way, with an effect consistent enough that it would be surprising for it to disappear. For example: creatine for muscle, caffeine for alertness.
Moderate. Real human evidence, with limitations — smaller trials, mixed results, or a genuine but modest effect. Worth taking seriously, worth holding loosely. For example: curcumin for joint pain.
Emerging. Early or preliminary human data, or a credible mechanism with only limited human evidence behind it. Promising, unproven. A good deal of the more interesting category sits here. For example: most of the longevity compounds.
Weak. Sparse, conflicting, or largely mechanistic and animal evidence. Often a compound whose reputation has run well ahead of its data. For example: ginkgo for memory, or the glucosamine and chondroitin joint pairings.
Unsupported. A claimed benefit that the evidence does not bear out, or that the evidence points against.
Grades are reviewed and revised as the evidence moves, so the examples above describe the picture as it stands rather than a fixed verdict.
We label grades in plain words and nothing else, no colours, no letter grades, no red-amber-green. Those devices tell you how to feel before you have read anything, and they tend to provoke defensiveness about your own stack. We would rather state the evidence and let you weigh it.
Where the honest grade is uncomfortable, we use it anyway. If a popular compound falls short of its reputation, the grade says so and the note beside it says why. A grading system that flatters everything is worth nothing; the value of a grade is precisely that it can be low. If anything we err towards caution — a compound has to earn its way up the scale, not down it.
The grades and the notes that accompany them are written by a trained doctor, working from the published human evidence rather than from marketing material. The database is built by hand, compound by compound, not scraped, not generated automatically, and each entry is reviewed before it ships. The same evidence ladder is applied across every compound, so a Moderate for one means what it means for another.