Paste your copy to catch the phrasing that most often gets ads rejected, before you burn a review cycle.
Most Meta ad rejections for copy fall into four buckets: implying personal attributes ("are you overweight?"), unapproved health claims, guaranteed income or results, and unverifiable absolutes like "#1" or "100% guaranteed". This checker scans your text for those patterns and explains how to rephrase each one.
This is a text heuristic covering the patterns that most often trigger rejection. It cannot see your images, landing page or account history, and it is not affiliated with Meta. Only Meta’s review decides.
Catch problems before the batch, not after the rejection
One policy problem in shared copy gets a whole batch rejected at once. Volume Creatives runs a pre-launch review of every ad it is about to create, so issues surface before anything publishes.
Paste the full ad text
Put the primary text, headline and description into the box together. Meta reviews all three, and a clean primary text does not save a headline that promises guaranteed results.
Read the flagged issues
Each hit names the category (personal attributes, health claim, income guarantee, superlative, platform name, urgency or sensational language), explains the policy behind it, and marks whether it is likely to block review.
Rewrite using the suggested fix
Apply the Try instead line, which usually means describing the product or the outcome rather than the reader, and dropping absolutes. Paste the new version back in until nothing red remains.
Check the rest of the ad yourself
The checker only reads text. Before launching, confirm the landing page makes the same claims, the imagery has no before and after comparison, and the account is not already under a restriction.
You enter
You get
“Are you in debt?” asks the reader about their finances.
“Earn $5,000 a month” is a specific earnings claim.
“results in 30 days” promises a timed outcome.
“guaranteed” on its own is an absolute Meta rarely accepts.
A rewrite that survives review describes the product and the typical experience rather than the reader and the outcome: a budgeting course that shows how members build an extra income stream from home. It says what you sell, it drops the number, and it removes every word that promises. It is also less clickable, which is the trade Meta is forcing, and the version that actually gets to spend.
| Category | Example that gets flagged | Fix |
|---|---|---|
| Personal attributes | Are you overweight? Struggling with anxiety? | Describe the product, not the reader: a plan that helps people lose weight. |
| Health claims | Cures diabetes. Reverses depression. | Drop cure, heal and reverse. Describe use and typical experience only. |
| Before and after | See the before and after. Results in 14 days. | Show the product in use; remove comparison imagery and timed outcomes. |
| Income guarantees | Earn $10k a month. Guaranteed returns. | Say what the product teaches or does; never state or imply earnings. |
| Superlatives | #1 in the world. 100% effective. Miracle. | Qualify or source the claim: rated 4.8 by 2,000 customers. |
| Platform names | The Facebook approved growth method. | Mention Meta or Instagram only factually and never imply endorsement. |
| Manufactured urgency | Act now. Only 3 left today. Last chance. | Use a real deadline you will honour or lead with the offer instead. |
| Sensational language | Doctors hate this. You won't believe. | State the specific claim you can back up on the landing page. |
These are the categories the checker looks for. Meta's full policy list is longer and also covers imagery, landing page content, restricted products and targeting, none of which a text scan can see.
Every ad goes through review before it delivers, and the first pass is automated. Models read the text, the image or video, the landing page and the account's history, and either approve, reject or route to a human. Because the first pass is a model, it is pattern driven: certain phrasings are strongly associated with past violations and get caught regardless of intent. That is why a supplement ad that says "reverses ageing" is rejected while a nearly identical ad that says "supports skin as you age" goes through. The words map to a pattern, not to a judgement about your product. This checker is built on the same idea from the other side: it looks for the phrasing that most reliably trips the review, so you can rewrite before submitting rather than after a rejection.
This is the single most common copy rejection and the least intuitive. Meta's policy forbids asserting or implying knowledge of a person's traits: their health, weight, finances, race, religion, sexual orientation, criminal record and so on. The test reviewers apply is whether the ad talks to the reader about themselves. "Tired of being broke?" fails. "A budgeting app for people who want to stop living paycheque to paycheque" passes, because it describes an audience rather than addressing a trait to the person reading it. The fix is almost always the same grammatical move: turn a second person question into a third person description of who the product is for.
The landing page is the next most common cause. Meta checks that the page makes the same claims as the ad and that it does not carry claims the ad avoided. A clean ad pointing at a page full of before and after photos and "guaranteed" badges will be rejected for the page, and Ads Manager will still label it as an ad problem. Imagery is the other one: comparison photos, close-ups of body parts, images of cash or medical devices all carry their own restrictions. And account history matters. A new account with no spend is reviewed more cautiously than one with a year of clean delivery, and an account with several recent rejections is reviewed more cautiously still.
Resubmitting the same ad unchanged, hoping a different reviewer sees it. Sometimes it works, and each time it does not it adds to the account's rejection count. Editing a live ad under review one word at a time, which restarts the clock with every save. Duplicating a rejected ad into a new campaign to dodge the flag, which Meta treats as circumvention. And appealing every rejection reflexively, which is fine when the rejection was wrong and counterproductive when it was right. The productive order is: read the stated policy, rewrite the flagged part properly, check the landing page, then resubmit once.
At one ad a day, a rejection costs a review cycle. At a batch of two hundred ads that share four primary texts, one phrase in one of those texts gets fifty ads rejected simultaneously, and the account health hit lands fifty times too. Checking the shared copy before it fans out is the highest leverage moment in the whole workflow. Pair this with the copy length checker so the approved text also survives truncation, use the naming generator so rejected variants are easy to find and pause together, and read why Facebook ads get rejected for the categories that go beyond text. When you are ready to launch, a bulk launcher with pre-launch guardrails runs this kind of check on every ad before anything publishes.
No, and no tool can promise that. Meta reviews your images, landing page, account history and ad copy together, and its review is largely automated with inconsistent outcomes even for identical text. This catches the text patterns that most often cause rejection; it cannot see the rest.
Meta prohibits copy that asserts or implies it knows something about the reader, especially health, finances, religion or other protected traits. "Are you struggling with debt?" implies knowledge; "a debt plan that works" does not. The rule catches questions aimed at the reader far more often than statements about the product, so rewriting second person questions into third person descriptions clears most of these.
Most ads clear within 24 hours, many in minutes. If an ad sits far longer than that, the usual causes are a restricted category, a new ad account with no history, or a landing page that does not match the ad. Editing an ad under review restarts the clock, so make all changes at once rather than one at a time.
Yes. Ads Manager and Account Quality both offer a Request review button, and appeals are worth filing when you believe the rejection was automated and wrong, because a human reviewer often overturns it. If the ad genuinely breaks a rule, appeal less and rewrite more: repeated rejections on one account raise the chance of a wider advertising restriction.
Because review is automated and probabilistic. The same text can pass one day and fail the next, particularly when paired with a different image or landing page, or after a policy update. Treat each rejection as a signal about the phrasing rather than a verdict on the whole concept, fix the flagged part, and resubmit.
Ads about social issues, elections or politics need authorisation and a paid for by disclaimer. Housing, employment and credit ads fall under Meta's special ad categories with restricted targeting. Alcohol, gambling, online pharmacies, cryptocurrency and some financial products have their own certification or country rules. Copy in those categories is reviewed more strictly and takes longer.
A single rejection does not. A pattern of them does, because Meta scores account health on rejection rate and on repeated violations of the same policy. Bulk launches make this worse in both directions: one bad line of shared copy can get fifty ads rejected at once, which is why checking the text before it fans out matters more than checking any individual ad.
Connect your ad account, drop in the creatives, and watch the batch land paused. Nothing spends until you say so.