Loading…
Paste the copy. Get the exact phrase that gets it rejected, plus the one-click fix that keeps the hook, in seconds — before you put spend behind it.
No integration, no upload, no account for the first one.
One creative, one market. Pick the vertical and the jurisdiction it runs in — the rules differ, and so does the verdict.
Every flagged phrase, the rule behind it, and the citation. Not a score: the sentence you have to change, and why.
A rewrite that keeps the hook, re-checked against the same rules before you ever see it.
A claim is allowed if it is true and you hold evidence for it. The argument is about whether your evidence is good enough. Substantiate it and you can run it.
A health claim is prohibited unless its wording is on a closed register (Reg (EC) 1924/2006, Art 10(1); CAP Code 15.1.1 for GB). Proof does not help. Being right does not help. If it is not on the list, it is not allowed.
And the safe-looking retreat is the trap. “Supports overall wellbeing” is only permitted next to a specific authorised claim (Art 10(3); CAP 15.2), so the vaguest line in the ad is the one that catches you — which is why this bites the careful operator, not the reckless one.
Paste your copy and see which lines survive in each market you sell into.
You're paying for the depth: vertical-specific policy packs, claim-level reasoning cited to the exact rule, the one-click fix, and rule-change tracking so nothing silently ages out. The part a generic checker can't fake.
A generic checker treats nutra like crypto. We don't. Each pack is built from the regulator surface that actually gets creatives rejected in that vertical.
The fear is not that a checker misses something. It is that it flags everything and leaves you with copy nobody clicks. So here is the other half of the judgement — each of these contains a word you would expect to catch, and each is silent for a different reason.
“Helps maintain cholesterol levels already within the normal range.”
Cholesterol is a disease risk factor, and naming one is normally a per-se drug claim. This phrasing is the exemption written into the rule itself.
FDA structure/function safe harbour — 65 FR 1000 (Jan 6, 2000)
“Clinically studied ingredients, at the dose used in the study.”
“Clinically” is a substantiation trigger. This survives it by being precise about what was studied and at what dose, which is what substantiation means.
“Results vary. Most customers report noticing a difference within eight weeks.”
A timeframe plus a result. Attributed and hedged, so it describes reports rather than promising an outcome.
We are not listing which patterns we run — that is the part a competitor would want. What is disclosed here is how the line is drawn.
The spend you can't recover. The winners you can't relaunch. The pixel you rebuild from zero, and the two weeks before it's smart again. Nobody publishes that number — you know yours. The first catch you make pays for years of this.
Methodology + reproducibility hash on the .
The same engine, the same packs, the same verdicts — behind an API key you mint yourself. Check before a creative reaches your uploader, or across every market you sell in, in one request.
POST /v1/check
X-API-Key: av_live_...
{ "text": "...",
"vertical": "nutra",
"jurisdictions": ["us", "eu"] }A general model will green-light this in a confident tone. Ours names the exact line and cites the law you can go and read for yourself.
Paste the one you're about to boost. One free check now, 8 more after sign-up, no card.
No card to start