Supplements · Netherlands & Belgium
Did anyone actually check what's in it?
Look up any supplement and see what a laboratory measured — next to what the label promised, and next to the document the number came from.
Press / anywhere to jump to search.
Start with what's in your cupboard
Sixteen categories, each with its own headline measurement — because milligrams of curcuminoids and billions of live bacteria are not the same question.
A tile takes you straight to the data. If you'd rather know what to look for first — what the category is, why people take it, and which markers matter — read the guide for each one.
What you get on every product
No letter grade, no star rating. The measurement, who took it, and the sentence it was read from — so you can check us.
Measured against the label
The bar shows what the lab found; the tick is what the pack claims. If nobody measured it, the bar stays empty — an untested product never reads as a pass.
Who paid for the test, in plain sight
An independent lab, a certification registry and a brand's own certificate are three different kinds of evidence. We colour them differently, list them separately, and never average them into one number.
The receipt is attached
Every figure carries the verbatim line it came from and where in the document it sits — PDF page, table name. Open the record and read it yourself.
The brand gets a right of reply
Any brand can respond to a result on its own products. Responses publish alongside the data, unedited.
How it works
We watch who publishes
Brand certificate pages, certification registries and regulator feeds are checked on their own schedule. When a new batch report appears, we fetch it and keep the original file.
We read the document, not around it
Each value is pulled together with the exact sentence it appeared in. Where a document is silent, the field stays empty — there is no way for us to fill in a guess.
We match it to the product on the shelf
By EAN first, then by name within a brand. Anything we aren't sure of goes to a human before it is published, because a real result on the wrong product is as bad as a made-up one.
Want the long version? Read the method and the rules we hold ourselves to.
"Is it any good?" is really six different questions
Potency is the obvious one, and it's rarely the one that catches a bad product. Each category fails in its own way, so each gets its own headline measurement — and every one of them still answers to the same test: did somebody who doesn't sell it ever measure it?
Every category
Is there as much in it as the label says?
Measured amount ÷ claimed amount — the % of claim. It's the one figure that compares across every product on the site — arithmetic, not a verdict.
Omega-3
Has the oil gone rancid?
Fish oil oxidises on the shelf, and rancid oil is the most common failure in the category. TOTOX puts a number on it — and almost nobody publishes it comparably.
Probiotics
Are the bacteria still alive?
A count taken at the factory is not the count in the capsule you swallow months later. We record what was counted, when, and on which lot.
Protein & plant powders
What came along with the crop?
Cadmium and lead concentrate in plant proteins and cocoa-flavoured blends. Protein content is also easy to overstate, so both get recorded.
Herbals
Is it the plant it says it is?
Curcumin and ashwagandha are extracts, so the active marker matters more than the milligrams of powder. Adulteration screens belong right next to the potency figure.
Vitamins & minerals
Is that 400 mg of magnesium, or of magnesium citrate?
They are very different amounts. We store the elemental figure and the measured dose, not the compound weight printed on the front of the tub.
Rules we hold ourselves to
These aren't values on a wall. They're constraints in the schema — the database physically cannot hold a number without the sentence it was read from.
01 Never invent a number
Every measurement carries the verbatim text it was extracted from and a locator — PDF page, table name, DOM path. If a value isn't in the document, the field stays empty. It is never inferred, and never carried over from a similar product.
02 Never merge self-reported with independent
A certificate the brand paid for and published is useful. It is not the same evidence as a lab test commissioned by someone else. Both are stored, always labelled, always displayed separately. No blended score.
03 Attribution errors count as fabrication
Attaching a real result to the wrong product is as damaging as making one up. Matches below our confidence threshold go to a human before they are ever published.
04 We report, we don't conclude
You'll find what a lab measured, who measured it, and when. You won't find us calling a brand a liar. Comparisons are factual and sourced, and every brand has a standing right of reply on its own page.
05 We don't touch paid test archives
Subscription test databases stay untouched. Where a consumer organisation has published a finding openly, we store the verdict, the citation and a link out — never a copy of their table.
06 The original document is kept forever
Every certificate we read is archived as fetched. If we get an extraction wrong, we can re-derive it; if a brand disputes a figure, we can show exactly what we read.
Two sources so far. IFOS publishes per-lot certificates for omega-3; ESN publishes per-batch lab reports for its protein powders, and is the only manufacturer in that category that does. ESN's labs forbid reproducing their reports, so those values are shown with the lab, lot and date and a link to the report — but the quoted line is withheld rather than copied.
Rule 01 is why the rest of the catalogue looks empty rather than full. A product nobody has tested shows no results, not a plausible-looking figure, and never reads as 100% of its claim. The emptiness is the honest state of the data, and it is what this site exists to fill in.
The same three steps, in detail
For readers who want to check our working rather than take it on trust.
Watch the sources
Brand certificate pages, certification registries and regulator feeds are polled on their own cadence. Change is detected by content hash, sitemap timestamp or registry diff — so a silently dead source raises an alert instead of quietly ageing.
Archive the original
The document is stored exactly as fetched, before anything reads it. Nothing downstream can overwrite the evidence.
Extract with spans, not guesses
Structured extraction pulls each value together with the verbatim string and its location in the document. Absent values come back empty — the schema has no way to express a guess.
Normalise and resolve
Units, per-serving versus per-capsule, and analyte naming are reconciled. The product is matched by EAN where one exists, then by exact name within a brand, then by similarity — with anything uncertain routed to human review.
Publish with provenance attached
The result appears with its lab, accreditation status, lot, sampling method, test date and a link to the source document. Provenance is not a detail view — it travels with the number.
Most brands publish nothing. That's a finding, not a gap.
We looked at twenty omega-3 brands on the Dutch and Belgian shelf. Three publish a certificate a buyer can actually read. Several advertise "independently tested" while publishing no test at all. Absence of evidence is information, so we show those products rather than hide them — you can filter the explorer to "no published data" and see how much of the shelf that is.
Three of twenty audited omega-3 brands publish readable lab certificates.
Figures from our own audit, 7 August 2026. That audit went deep on omega-3 as the pilot category; the same source walk is being repeated category by category. Product data on this site is real and partial.
Publishes per-batch certificates
The transparent minority
A handful of brands put real certificates online — per batch, matched to the number printed on the pack. In the audit these were Testa, Nutribites and Nordic Naturals.
Claims without documents
"Independently tested"
The most common posture: strong purity language, specific-sounding limits, and nothing published to check them against.
Silent
Private label and mass market
Drugstore own-brands publish no test data at all, and several actively block automated access to their own product pages.
Registries
Where the real numbers live
Certification programmes are the richest legitimate source. Most publish only a pass/fail badge, but IFOS publishes per-batch reports with actual measured values.
About
Why we're called Assayer
An assayer measures the purity of a metal and reports the number. Sworn to the standard rather than to the seller — no stake in the goods, and no opinion on whether you should buy them.
The oldest version of this job
An assay office does not tell you the ring is beautiful. It tells you it is 585 parts gold per thousand, strikes that on the metal, and leaves the decision to you. The value of the office comes entirely from the fact that it does not sell rings.
That is the whole job description here. We don't run tests. We find the tests that already exist, read them the same way every time, and put each number next to the document it came from.
The name has a second owner
In October 1623 Galileo published Il Saggiatore — The Assayer. It was a reply to a rival who had published his own attack under the title Libra, the balance. Galileo's answer was a pun with an argument inside it: a libra is the scales you weigh vegetables on, and what these claims deserved was the assayer's balance — the finest instrument of the age, used to weigh precious metal.
Four centuries later that book is read as one of the founding statements of the scientific method: test the claim against the measurement, instead of accepting it from whoever asserted it most confidently.
The supplement shelf is still weighed on the grocer's scales. "Independently tested." "Pharmaceutical grade." "Third-party quality assured." We put the certificate on the finer instrument and show you the reading.
What the name promises — and what it doesn't
An assayer reports. An assayer does not certify, rank, endorse or recommend. You'll find what a laboratory measured, who measured it and when. You won't find us calling a brand a liar or crowning a winner, and there is no badge here for anyone to buy.
That constraint sits in the name deliberately. A promise you have to repeat every time someone asks what you do is a promise that's harder to quietly drop.
Questions people ask us first
Do you test the products yourselves?
No. We are not a laboratory and we don't commission tests. Everything here was already published by someone else — a brand, a certification programme, a regulator — and our job is to find it, read it consistently, and put it next to the product you're holding.
So which supplement is the best one?
We won't tell you that, and it's deliberate. A single grade quietly blends a lab test somebody paid for with a certificate the brand published about itself, and once it's one number you can't see which is which. We show you the measurement, who produced it, and when — the ranking is yours to make.
Does a brand pay to appear here, or to rank higher?
No. There is nothing to buy. Brands that publish good data show up with good data attached; brands that publish nothing show up with an empty record, and that's a fact about the brand rather than a punishment.
Why do so many products say "no published test data"?
Because most brands genuinely publish none. That row is not a failing grade — it means nobody has made a result public that we could check, which is itself the most useful thing we can tell you about the shelf.
A number on my product looks wrong. What now?
Tell us and we'll re-derive it. We keep every source document exactly as we fetched it, so we can show you the sentence we read and correct the record if we got it wrong. If you're the brand, you can also file a response and we publish it next to the result, unedited.
Is this medical advice?
No. Nothing here tells you what to take or whether to take anything at all. It tells you what a laboratory measured in a particular lot of a particular product on a particular date.
Look up what you're already taking
Filter by category, potency, oxidation, evidence type and certification. Open any product for the full analyte table with its sources attached.
Explore the data →Free · no account · nothing to buy