OnlyFans has never published a list of restricted words, and it never tells you which word caused a rejection. So here is the honest version: the ten categories the filter actually enforces, where it runs, and the one place it fails silently and quietly costs you sales.
Last updated August 2026
Paste your text. It is scanned in your browser and never sent anywhere.
Result
Nothing checked yet. The checker covers the categories most likely to be filtered by accident, so treat a clean result as encouraging rather than as a guarantee.
No flagged terms
Nothing in the categories below. This is not an official check, and OnlyFans changes the filter without announcement, so if the text still refuses to save the cause is usually a phrase rather than a single word.
Unofficial and not affiliated with OnlyFans. There is no published list to check against, so this reflects the categories corroborated across 2026 creator-facing sources. The binding version is the Acceptable Use Policy on your own account.
OnlyFans filters restricted words in ten categories: age and youth language, consent and coercion, family relation words, violence and injury, bodily waste, drugs and intoxication, sex work and meeting up, off-platform payment apps, competing platform names, and slurs. The filter runs on your display name, bio, post captions, direct messages and mass messages. There is no official published list, and the platform does not tell you which word failed.
The part almost every guide gets wrong is which words actually cost working creators money. It is not the obviously prohibited vocabulary, because nobody was going to type that anyway. It is the innocuous words: teen, young, drunk, meet up, in person, PayPal, Venmo, Fansly. Those appear in perfectly ordinary captions written by creators who have no idea they are anywhere near a policy line.
And one behaviour is worth knowing before anything else on this page. In a bio or a caption, a filtered word is obvious: the field refuses to save. In a direct message it is not. The message appears in your own thread as though it sent, the subscriber never receives it, and you get no error at all. Creators lose real money to that gap for months while concluding their fans have gone quiet.
Grouped by category rather than printed as a single alphabetical list, for a reason. No canonical list exists to reprint, the filter changes without notice, and a category tells you how to rewrite a sentence while a word list only tells you what failed. The examples are representative, not exhaustive.
| Category | Example terms | Why it is filtered | What to write instead |
|---|---|---|---|
| Age and youth language | teen, young, little girl, schoolgirl, barely legal, jailbait | The single most common accidental trip. The filter cannot tell a of-age roleplay reference from a real one, so it treats the vocabulary itself as the risk. This category is enforced hardest and appealed least successfully. | Describe an adult in adult terms, or say nothing about age at all. There is no phrasing of "teen" that reads as safe to an automated filter. |
| Consent and coercion | forced, unwilling, passed out, unconscious, drugged, chloroform, kidnap | Anything implying a person cannot agree to what is happening. Fantasy framing does not exempt it, because the filter reads the words and not the framing around them. | Rewrite the scenario as enthusiastic and explicit about consent. If the concept only works without consent, it does not work on the platform. |
| Family and relation words | incest and step-relation phrasing used to describe a scene | Family roleplay sits in the prohibited categories rather than the grey area. Note that ordinary affectionate nicknames are not the issue here; describing a scene as a family relationship is. | Drop the relationship label. The same content described without it is usually fine. |
| Violence and injury | torture, blood, snuff, strangle, choke, asphyxiation | Real or implied harm, and breath-play vocabulary in particular, which reads to a filter as injury risk regardless of how the scene was actually conducted. | There is no safe synonym for this one. If the content depicts it, the wording is not the problem you need to solve. |
| Bodily waste | pee, peeing, piss, pissing, scat, vomit, vomiting | The clearest example of a rule that comes from outside the platform. Payment processors decline to fund this category, so it is filtered whether or not it is legal where you live. | Nothing. This category is prohibited outright rather than merely filtered, so a euphemism does not make the content permitted. |
| Drugs and intoxication | cocaine, heroin, meth, weed, drunk, high, wasted | Illegal substances, paraphernalia, and any suggestion that someone is too intoxicated to consent. "Drunk" catches creators who meant it as a throwaway joke in a caption. | Cut the reference. A caption rarely needs it, and it is a common source of a warning on an otherwise clean account. |
| Sex work and meeting up | escort, escorting, prostitute, hooker, meetup, fly me out, in person, rates | Anything that reads as arranging a paid physical encounter. This is a hard line for processors and it catches innocent phrasing: a caption offering to "meet up" for a giveaway can trip it. | Keep everything you offer explicitly on-platform and digital. Avoid "meet", "in person" and "rates" even in unrelated sentences. |
| Off-platform payment | PayPal, Venmo, CashApp, Zelle, Bitcoin, wire | Naming an external payment method looks like routing money around the platform, which is both a terms breach and a processor concern. Filtered in bios, captions and DMs alike. | Use the on-platform tip and paid-message tools. If you want fans buying things off-platform, use a wishlist or storefront link rather than a payment app name. |
| Competing platforms | Fansly, ManyVids, Fanvue, JustForFans, Chaturbate | Naming a rival is treated as diverting subscribers you acquired on the platform. This is the one category that is commercial rather than safety-driven, and it is enforced quietly. | Point fans to a single link hub you control and let that page do the routing, instead of naming platforms in a caption or DM. |
| Hate speech and slurs | racial, ethnic and identity-based slurs | Filtered platform-wide and not specific to adult content. Also the category most likely to draw a permanent action on a first offence rather than a warning. | Nothing to rewrite. This one is straightforward. |
Two rows deserve a second look because they catch people who are not doing anything remotely against the rules. "Sex work and meeting up" flags ordinary words like "meet up" and "in person", so a caption about a giveaway or a convention appearance can fail. "Off-platform payment" flags the app names themselves, which means a message explaining to a fan that you do not use Venmo will be blocked for containing the word Venmo.
No, and this is worth being blunt about because so much of the internet implies otherwise. OnlyFans does not publish a restricted-words list, in the Acceptable Use Policy or anywhere else, and the platform does not identify which word caused a rejection when one happens.
You can see the consequence on any search results page for this topic. In the same week you will find sites confidently offering "the 206 banned words", "200+ restricted terms", "122 restricted words" and "100 banned terms". Those are four different numbers for one list, and they disagree because every one of them was reconstructed by somebody typing phrases into the platform and writing down what failed. That is genuinely useful research. It is not an official document, and treating it as one is how creators end up trusting a list that describes a filter from eighteen months ago.
The filter also moves. When a payment processor tightens its conditions, the wording rules tighten with it, and phrasing that published fine last year gets actioned this year with no announcement. A category-level understanding survives that. A memorized word list does not.
Every text field on the platform is scanned, but the failure does not look the same in each one. The difference between the top three rows and the bottom two is the difference between a mild irritation and money you never knew you lost.
| Where | How it fails | What you actually see |
|---|---|---|
| Display name and username | Rejected on save | The field refuses to update and shows an error. Annoying, but harmless: you find out immediately and nothing has been published. |
| Bio | Rejected on save | Same behaviour. A red outline on the text box and the change does not stick. This is the safest place to discover a flagged word. |
| Post caption | Rejected on save | The post will not publish while the caption contains a filtered term. Creators often assume the upload failed and retry the file, when the caption was the problem. |
| Direct message | Sends on your side, never arrives | The one that quietly costs money. The message appears in your thread as though it went, and the subscriber never receives it. You get no error, so you conclude the fan ignored you. |
| Mass message | Sends on your side, never arrives | Identical failure mode at scale. A promotional blast with one filtered word in it can reach nobody while your dashboard shows it as sent. |
The silent DM failure is the reason this page exists. A creator sends a paid message to two hundred subscribers, sees it in every thread, and gets four replies instead of forty. The obvious conclusion is that the offer was wrong or the list has gone cold. The actual cause is one word. If your reply rate fell off a cliff without an obvious reason, re-read your recent mass message templates before you rewrite your whole offer.
One filtered word, once, is a non-event. The field does not save, or the message does not arrive, and nothing else happens. Nobody is suspended for typing "teen" into a caption box on a Tuesday.
What escalates is the pattern. Repeated attempts at the same blocked wording read as somebody probing for a way through rather than somebody who made a typo, and that flags the account for manual review. From there the outcome depends entirely on which category the wording sits in. Payment-app names and competing platform names tend to produce a warning. Age and consent language can produce a suspension on a much shorter fuse, because those are the categories the whole filter exists to protect.
This is also where the appeal advice most creators receive goes wrong. If your account is actioned over wording, the appeal that works quotes the exact caption and asks which clause it breached. The appeal that fails argues that the content was legal and consensual, which is true and irrelevant, because the rule being enforced was never about legality. The full breakdown of which enforcement categories an appeal can realistically reverse is on our OnlyFans account banned guide.
The list looks arbitrary until you know where it comes from, and then it stops looking arbitrary almost entirely. A large share of it does not originate with OnlyFans at all. The platform sits downstream of card networks and payment processors, and those companies set conditions on what they are willing to fund. Those conditions become clauses in the Acceptable Use Policy, and the clauses become an automated wording filter.
Read the categories again with that in mind and they resolve into two groups. Age, consent, violence, bodily waste and drugs are all things a processor has declined to fund, which is why they are enforced regardless of what is legal where you live. Payment apps and competing platform names are commercial: they concern money leaving the platform. The only category that is neither is slurs, which is a straightforward content rule that every mainstream platform has.
That framing is the practical difference between guessing and knowing. If you can tell which of the two groups a phrase belongs to, you can usually tell whether it is worth rewriting or whether the underlying content is the problem. It also tells you when to re-audit: not on a schedule, but whenever the payments industry tightens, which is when the wording rules quietly follow.
Four habits cover almost all of it, and none of them make your writing blander.
1. Test in a caption box, not in a DM. If you are unsure about a phrase, paste it into a post caption first. That field tells you immediately whether it saves. A DM will not, and by the time you notice, the message has already failed to reach a paying subscriber.
2. Never negate a flagged word. The filter matches the term, not the sentence around it. "I don't do meetups" and "please don't ask me to meet up" both contain the flagged phrase and both fail. Say what you do offer rather than naming what you do not.
3. Route traffic through one link you own. This solves the payment-app and competing-platform categories in a single move. Instead of naming Venmo or Fansly anywhere, point everything at a hub you control and let that page do the routing. It is better for you regardless, since it is the one asset a suspension cannot take. There is more on structuring that in our guide to an OnlyFans link in bio.
4. Re-read the templates you reuse. Saved captions and welcome sequences are where old wording survives, because nobody re-reads a template that worked. A phrase that was fine when you wrote it can be filtered now. It is worth running your OnlyFans captions and your welcome message through the checker above every few months, and the same goes for the opening lines in your DM strategy.
OnlyFans filters words in ten broad categories: age and youth language, consent and coercion terms, family relation words, violence and injury, bodily waste, drugs and intoxication, sex work and meeting up, off-platform payment apps, competing platform names, and slurs. The words that catch working creators are ordinary ones like teen, young, drunk, meetup, PayPal and Fansly rather than the obviously prohibited terms.
No. OnlyFans does not publish a restricted-words list anywhere, and it does not tell you which word triggered a rejection. Every list you find online is reconstructed by creators and agencies from trial and error, which is why the published counts disagree so widely: the same week you can find sites claiming 100, 122, 200 and 206 terms.
The most common cause is a filtered word in the message itself. In direct messages the filter fails silently: the message appears in your thread as sent, but it is never delivered to the subscriber and you receive no error. Rewrite the message without the flagged term and send it again rather than assuming the fan ignored you.
A single blocked word usually just fails to save, with no further consequence. Repeated attempts are what escalate: the pattern flags the account for manual review, which can produce a warning, a temporary suspension, or a permanent ban depending on which category the wording falls into. Age and consent language escalates fastest.
No. Competing platform names are filtered because naming a rival is treated as diverting subscribers you acquired on OnlyFans. This applies in bios, captions and messages. If you want fans to find you elsewhere, route them through a single link hub you control rather than naming the platform in text.
Yes, as words as well as as payment routes. Naming an external payment app in a bio, caption or DM reads as an attempt to move money off-platform, which breaches the terms and creates a processor problem. Use the built-in tip and paid-message tools, or a wishlist link, instead.
An automated filter scans the text of every message before it is delivered, in the same way it scans bios and captions. That is not a human reading your conversations, but it does mean message content is processed, and it is why a filtered word stops a DM from reaching the person you sent it to.
Because none of them are official. Each list is somebody reconstructing the filter by testing phrases and recording what failed, at a particular moment. The filter also changes without announcement when payment processors tighten their conditions, so a list published a year ago describes a filter that no longer exists.
Wording is one of several ways an account gets into trouble by accident. The others worth auditing while you are here are covered on OnlyFans verification and OnlyFans chargebacks, which between them account for most of the suspensions that had nothing to do with the content itself.
Knowing the categories is the easy half. The hard half is writing hundreds of captions and messages a week that are compelling, on-brand and clean, and then noticing the day a phrase that always worked quietly stops being delivered.
That is the part we run for creators, along with the promotion accounts that tend to get actioned before the OnlyFans profile ever does. Free, confidential application, and we reply within 24 hours whether it is a fit or not.
Tell us where you are now and we will tell you honestly what we can do.
Captions, bios, DM scripts, tip menus, pricing and more, generated in seconds. No card needed.