Acceptable Use Policy

Content involving minors: zero tolerance

Nothing on this service may sexualise a minor, or anything presented as one. This rule has no exceptions: no artistic exception, no roleplay exception, no fiction exception. That a character is invented and machine-generated changes nothing about it.

All of the following are prohibited, in prompts and in output alike:

• text of any kind — a prompt, a chat message, a character name, biography or scenario — that places a minor in a sexual context, or that asks for one;
• any character stated, implied or numerically set to be under 18. Every character on this service is represented as 18 or older, and a character you create must state an age of 18 or above;
• wording that stands in for an age without naming one: describing a body, proportions or a voice as a child's, using childhood forms of address, or pairing sexual content with school uniforms, classrooms, playgrounds, nurseries, cots or toys;
• uploading a photograph of a minor for any purpose whatsoever, including as a source image for generation;
• sharing, requesting or linking to such material anywhere on the service, including in comments;
• any attempt to get around the filters — coded words, deliberate misspellings, another language, splitting a request across several messages, walking a character's stated age downwards, or asking a character to describe itself in those terms. The attempt is the violation. It does not have to succeed.

The mirror of this rule is who may be here at all. The service is for adults. You must be 18 years of age or older to use it, or older where the law that applies to you sets a higher age of majority. The 18+ confirmation you clicked is your own declaration and nothing more — we do not check documents, faces or payment records, and we do not claim to. It rests on you telling the truth. Do not let anyone under that age use your account, use your device while you are signed in, or see content taken from this service.

Real, identifiable people

Sexual or intimate content about a real, identifiable person, produced without that person's consent, is prohibited. It makes no difference whether the person is famous, whether the result is convincing, or whether you say openly that it is generated.

• uploading a photograph of another person, or any photograph you do not have the right to use, as a source for generation;
• “undressing” a photograph: generating a nude, partially nude or sexual image from a clothed one;
• face swaps, deepfakes and voice clones of a real person, in a sexual context or in any other;
• building a character on a real person's name, face, voice or life story, and in particular publishing such a character;
• using another person's photograph as your profile photo or as a character's image;
• generated sexual content about someone you know personally — an ex-partner, a colleague, a classmate, a neighbour. This is the most common form of this abuse and among the most damaging, and it is treated as a serious breach even when it is never shown to anyone else.

A person depicted in generated material may have it removed regardless of who produced it, and so may anyone acting for them. Use Content Removal & DMCA, which handles content published without the consent of the person depicted within 24 hours rather than in the general queue.

Sexual content prohibited outright

The categories below are prohibited by the rules of the card schemes we accept payment under, and independently by us. They are prohibited in text and in generated media, and they are prohibited in fiction: a scene between two invented characters is still a scene we host and still one we are answerable for.

• sexual activity presented as non-consensual — rape, sexual assault, coercion, blackmail for sex, or a partner who has refused;
• sex with a person who cannot consent: unconscious, asleep, drugged, heavily intoxicated, hypnotised, or restrained without having agreed to it;
• incest, and sexual content between characters presented as close family;
• bestiality, and sexual content involving animals or characters presented as animals;
• sexual violence: torture, mutilation, strangulation, or anything that sexualises serious injury, death or a corpse;
• trafficking in persons, sexual slavery, forced prostitution, and the advertising, arranging or solicitation of prostitution or escort services;
• sexualising a person's captivity, kidnapping, or someone's custodial or institutional power over them;
• content that sexualises or celebrates violence against a group of people, or that pairs sexual content with hatred of a protected characteristic;
• instructions, encouragement or practical help for doing any of the above in the real world.

Framing does not create an exception. Labelling a scene as consensual roleplay, as fantasy, or as agreed in advance between the characters does not change what the content depicts. If it reads as sexual violence, it is prohibited.

Passing generated content off as real

Everything this service produces is synthetic — our §2257 Exemption Statement states the same fact from the other side. Presenting the output as a photograph, a recording or a video of a real person or a real event is prohibited, both here and anywhere else you take it.

• publishing or sending a generated image, video or voice clip as a genuine photograph or recording;
• claiming that a character is a real person, that a real person appears in generated media, or that a real person consented to it;
• impersonating anyone: another user, a public figure, a business, or our own staff, support or moderators;
• removing or obscuring any marking, label or metadata that identifies media as generated, where one is present;
• using generated media to harass, threaten, humiliate, defame or damage the reputation of anyone;
• extortion — threatening to publish generated sexual imagery of someone unless they pay, send images, or do anything else. This is a criminal offence in most jurisdictions and we treat it as one.

Comments, nicknames and published characters

Some of what you make here other people can see: a character you mark public, your nickname and profile photo, and comments under public characters. A character you publish waits for review before it appears in the catalogue, and it can be removed after that review as well. Everything above applies to these surfaces, and so does the following.

• harassing, threatening or persistently targeting another user, including following them from thread to thread;
• attacks on a group of people by race, ethnicity, nationality, religion, disability, sex, gender identity or sexual orientation;
• publishing another person's private information — address, telephone number, workplace, private messages or photographs;
• spam, advertising, referral links, and links to malware or to other adult services;
• content that promotes suicide, self-harm or eating disorders, or that glorifies serious violence;
• offering or seeking anything illegal: drugs, weapons, stolen data or stolen credentials, sexual services;
• a nickname, profile photo or character image that impersonates a real person or breaches any other rule in this policy.

Access, automation and commercial exploitation

Your access is personal to you. The Terms of Service license it for your own use; it is not stock you can pass on. Prohibited:

• reselling, renting or sublicensing access, or offering the service on to others as your own;
• sharing your account, password or session token with anyone, or letting more than one person use a single account;
• opening further accounts to claim a free allowance, a promotional code or a reward again, or to return after a termination;
• automated access of any kind: bots, scripts and crawlers, or scraping characters, comments, media or catalogue data;
• reverse engineering or decompiling any part of the service, probing it for vulnerabilities, or attempting to extract model weights, system prompts or training data;
• using output from this service to train, fine-tune or benchmark a competing model, or to build a competing service;
• circumventing metering or payment: bypassing the free-message allowance, the paywall or gem accounting, or interfering with how a purchase is recorded;
• selling generated media as authentic photography, as stock imagery of a real model, or with any claim that a real person appears in it or endorses something;
• selling, trading or transferring gems, characters or accounts. Gems are a virtual in-app currency, are not money, and do not move between accounts.

Payments

Checkout is hosted by our payment provider on its own form, so card details never reach our servers. The following are breaches of this policy, and the first two are criminal offences in most places as well:

• paying with a card, account or wallet you are not authorised to use;
• using card details you know to be stolen, or any payment you know to be fraudulent;
• disputing a purchase with your bank that you received and used, without ever having raised it with us;
• abusing promotional codes, rewards or referral mechanics, whether through additional accounts or otherwise.

We ask you to come to us first because we can usually settle it faster than a bank can: you have 60 days to raise a billing problem with us, and the Refund & Cancellation Policy sets out how. Where a payment is reversed we may remove the gems or subscription time it bought, and we may suspend the account while the dispute is open. Asking us first is a request and not a condition: the right to take a charge to your bank belongs to the cardholder, and nothing in this policy removes or delays it.

How this is enforced, and where the filters fail

Two automated layers sit between a request and a result. It is worth stating exactly what each one does, because neither is a guarantee and we would rather you knew that than found out.

Before generation

The text you wrote — the prompt and the fields a scene is composed from — is screened automatically before a generation runs. A request that fails the screen is refused, and nothing is generated.

After generation

Generated images pass an automated nudity and safety classifier on the generation side, independently of the screen on the text. An image it rejects is not delivered, and the gems reserved for it are released.

After a report

A person reviews reported material and the account behind it. That is where a decision to remove content or terminate an account is made; the automated layers do not make it.

Both automated layers are imperfect, in both directions. They refuse requests that this policy permits, and they will sometimes let through something they should have stopped. Neither outcome changes anything: content that breaches these rules is a breach whether or not a filter noticed, and a filter having let something through is not permission. We also do not publish the filters' rules in detail, because a precise list of triggers is a manual for evading them.

What happens when these rules are broken

Enforcement is carried out by [LEGAL ENTITY NAME], the operator of the service. There is no fixed number of warnings. What we do depends on what the breach was and whether it has happened before, and it can be any of:

• removal of the content, including any character, comment, image or clip involved;
• loss of access to a feature;
• suspension of the account while a report is reviewed;
• immediate and permanent termination, without refund, for a serious breach or a repeated one.

Termination for a breach of this policy forfeits any unspent gems and any remaining subscription time, and no refund is due — the Terms of Service sets that out, and the Refund & Cancellation Policy covers ordinary cancellations, which are a different thing. Gems you have already spent are gone in every case: they are consumed the moment the action runs.

For material that sexually exploits a minor there is one outcome and it is not weighed against anything: the account is terminated immediately, the material and the account's records are preserved rather than deleted, and the matter is reported to law enforcement and to the competent child-protection reporting body — where a law that applies to us requires that report, and as our own policy in every other case. A termination on that ground is not reversed on request, and we do not discuss the content of such a report with the account holder. We may also contact law enforcement over a credible threat of violence or an attempt at extortion.

If you believe an enforcement decision was wrong, say so through the Complaints Policy. You will have a reply from a person within 48 hours and an outcome within 30 days, and if that answer does not satisfy you the complaint is escalated to somebody who did not take the decision. No enforcement decision is outside that route, including one taken under the first section of this policy — what a complaint about that one can establish is that we acted against the wrong account, not that the rule should have been applied more leniently.

Reporting a violation

If something here breaches this policy, report it. You do not need an account, and you do not need to identify yourself.

Material — an image, video, clip or published character

Use Content Removal & DMCA. Give the address or identifier of the material, what is wrong with it, and a way to reach you. General notices are answered within 7 days.

A minor, or content published without the consent of the person depicted

Write to [email protected] and put URGENT in the subject line. These reports leave the general queue and are handled within 24 hours, from anyone, without proof of identity.

Conduct — another user, a comment, harassment

Use the Complaints Policy route, or write to [email protected]. A person replies within 48 hours.

If anyone is in immediate danger, contact your local emergency services first and then tell us. We are not an emergency service and we cannot act at that speed.

Changes to this policy

This policy changes when the product changes or when a rule binding us changes. The current edition is always at this address, and its date is at the top of the page. We do not send an email when it changes and there is no notification list, so the date is what to check. When a change is material we raise the edition of our documents, and the 18+ confirmation is then asked again the next time you open the site.

Questions about this policy go to [email protected]. Anything that needs acting on should go through the reporting routes above instead, where it will be seen sooner.