Content & Acceptable Use Policy

Last updated: 15 July 2026

Vellux is operated by Dennis Müller, c/o Anibarro, Unit #152, Route de Saint-Cergue 24Bis, 1260 Nyon, Switzerland ("Vellux", "we", "us", "the operator"). This Content & Acceptable Use Policy ("Policy") is part of, and incorporated by reference into, our Terms of Service. It applies to everything you do on the Vellux platform at vellux.ai (the "Service"): the prompts and search terms you enter, the reference images, masks, videos, training photos, and LoRA files you upload ("Content"), and the images and video the Service generates for you ("Outputs"). Capitalised terms not defined here have the meaning given in the Terms of Service.

By using the Service you agree to this Policy. If you do not accept it, do not use Vellux.

1. Our stance: uncensored, not unmoderated

Vellux is an uncensored creative tool for adults. You must be at least 18 years old and of legal age in your country of residence to hold an account (see the age and territorial-availability rules in Section 2). Within that boundary we do not police taste, kink, aesthetics, or fantasy. Explicit adult (18+) content is permitted. We do not apply subjective "community standards," and we do not decide what is or is not tasteful on your behalf.

You are solely and fully responsible for the Content you supply and the Outputs you generate. Your prompts, your uploads, your models, and your results are yours to answer for. We provide a probabilistic generation tool; we do not vet the artistic or moral merit of what you make with it.

"Uncensored" is not the same as "lawless." It does not mean that illegal material is tolerated, and it does not mean the Service goes unmoderated for illegality. The prohibitions in this Policy are hard limits that override the permissive stance in every case. Where local law where you sit would technically permit something we prohibit here, our prohibition still governs your use of Vellux — these are contractual rules of the platform, and several of them exist because our payment partners require them regardless of local legality (see Section 7).

The Vellux Gallery is private by default. Your Outputs and uploads are stored in a per-account store served only through short-lived, presigned links; there is no public feed, profile, or sharing surface. Nothing you generate is published to a public audience on Vellux. What you then do with your Outputs off-platform is your responsibility, subject to the disclosure duties in Section 9.

2. Age and territorial availability

Vellux is strictly for adults. You must be at least 18 years old — and of the age of majority in your country of residence — to hold an account. You confirm this by an explicit, recorded attestation when you register, and email verification is mandatory for every email-and-password account (Fanvue sign-in relies on the identity Fanvue confirms).

Some jurisdictions require services that make adult content available to verify users' ages with methods stronger than a self-declaration — for example the United Kingdom (Online Safety Act 2023, "highly effective age assurance"), France (SREN law and the Arcom age-verification standard), Germany (JMStV closed-user-group standard), Italy (the AGCOM age-verification regime, Resolution 96/25/CONS) and Australia (the Online Safety Act 2021 age-assurance industry codes, which expressly extend to generative-AI services). Vellux is not offered in those jurisdictions, with one exception: in the United Kingdom and Australia, the Service is available solely to Fanvue creators signing in with their Fanvue account. Fanvue verifies every creator's identity and age — government-ID plus liveness verification — as part of its creator onboarding, and we rely on that third-party verification as the age assurance for exactly this group of users, a form of reliance both regimes' frameworks accommodate. All other access from the listed jurisdictions is blocked: we do not target those jurisdictions, we do not market to them, and we block access from them. Do not attempt to access the Service from a jurisdiction where it is not offered to you, or to circumvent a territorial block: doing so is a breach of this Policy, and any such access happens on your sole initiative, not because the Service is directed at you. If we choose to serve a restricted jurisdiction more broadly, we will first deploy an accredited age-verification method that meets its standard.

Where a payment provider's processing of a successful transaction gives us an age-relevant signal, we may rely on it in addition to your attestation. We may ask for further proof of age at any time, and we will suspend an account we reasonably believe is held by a person under 18. Providing false age information is a breach of this Policy.

3. Absolute prohibition: no sexualisation of minors, ever

There is zero tolerance, and no exception of any kind, for any sexual or sexualised depiction of a minor. This is the single most important rule on Vellux and it overrides everything else in this Policy and in the Terms.

This prohibition covers material that is real, fictional, drawn, cartoon, CGI, or AI-generated, including so-called "pseudo-photographs" and any wholly synthetic image. "It is not a real child" and "it is only synthetic" are not defences and will not be treated as such. The prohibition specifically includes:

• any depiction of a person under 18 (or who appears to be under 18) in a sexual, sexualised, or explicit context — we apply an "appears under 18" standard, so ambiguous, "aged-up," or age-uncertain presentation of a childlike subject is prohibited;
• "age-play," "teen"-coded, "schoolgirl/schoolboy," "loli/shota," or comparable youth-coded sexual content and search or prompt terminology;
• uploading, training on, or generating from images of a real minor for any sexual purpose whatsoever;
• using the character/LoRA feature (Section 6) to produce a minor's likeness in any sexual context.

You may not use Vellux to create, request, upload, or store child sexual abuse material ("CSAM") in any form. This mirrors the criminal law of the jurisdictions we operate under and serve, none of which recognise a "synthetic child" exemption — including Swiss Criminal Code Art. 197, the UK Protection of Children Act 1978 (pseudo-photographs), Coroners and Justice Act 2009 s. 62 (non-photographic prohibited images), the Crime and Policing Act 2026 (CSA-image-generator offence), and EU Directive 2011/93/EU.

Where we identify apparent CSAM we will: remove and block it immediately; preserve the material and associated account and log data as evidence to the extent legally required; terminate the account; and report to the competent authorities and reporting bodies (including fedpol in Switzerland, NCMEC's CyberTipline, and the UK NCA where applicable). There is no notice, warning, appeal, or counter-notice process for CSAM.

4. Real, identifiable people and consent

Any Content or Output that depicts a real, identifiable person requires that you hold that person's documented, express, written consent to be depicted in the way you are depicting them. For your own face and likeness, your own consent suffices.

The following are prohibited without the depicted person's documented written consent:

• generating, uploading, or storing images or video of an identifiable real person;
• any sexual, intimate, nude, or otherwise degrading depiction of a real, identifiable person — the creation of non-consensual intimate imagery ("NCII") and sexual "deepfakes" of real adults is prohibited outright, whether or not you intend to share it;
• using the face or likeness of a public figure, celebrity, influencer, politician, or any other identifiable person — public status is not consent, and impersonation, "look-alike," or "deepfake" use of a well-known person is prohibited;
• producing look-alikes or impersonations designed to pass an Output off as a genuine depiction of a specific real person.

This reflects, among others, the UK Data (Use and Access) Act 2025 s. 138 amending the Sexual Offences Act 2003 (non-consensual creation of intimate images), EU Directive (EU) 2024/1385 Art. 5, Swiss Criminal Code Art. 179quater and Swiss Civil Code Art. 28 (protection of personality), and the US TAKE IT DOWN Act (48-hour NCII removal). Sexual content involving a real, identifiable adult is permitted only where you hold and can produce that person's documented, express, written consent.

Any person depicted without their consent — or their representative — may request removal at any time through our Complaints, Reports & Takedowns process. Valid non-consensual-intimate-image requests are prioritised on a 24-hour fast track.

5. Uploads: you must own or be licensed for what you supply

When you upload any Content (a reference image, mask, video, training photo, or a ready-made LoRA file) you warrant that:

• you own it, or you hold all rights and licences needed to upload it and to have Vellux process it to generate Outputs;
• it does not infringe any third party's copyright, trademark, or other intellectual-property right, and is not stolen, leaked, or scraped without authorisation;
• it contains no CSAM and no minor in any sexual context (Section 3);
• for every real, identifiable person in it, you hold that person's documented written consent (Section 4);
• it contains nothing otherwise unlawful or prohibited by this Policy.

Do not upload content you found online, took from someone else's account, or obtained without the right to use it.

6. Characters, LoRA training and real-person likeness

Vellux's character features are built for synthetic personas — invented people who do not exist. The character builder deliberately blends several uploaded face and body datasets into a new, unique likeness and is not designed to return a model of any single input face. Two further paths — training an additional model for an existing character from a single uploaded dataset, and uploading a ready-made LoRA file — can produce a model of a specific likeness. Because a likeness model can reproduce a specific face on demand, these are the highest-risk features on the platform and carry the strictest rules; the categories and attestation below apply to every path.

Every training dataset, and every uploaded LoRA file, must fall into exactly one of three categories:

• a wholly synthetic persona — every image is AI-generated and depicts no real, identifiable person (this is what the feature is for);
• yourself — your own face and likeness; or
• a real person whose documented, express, written consent you hold for that specific use, including any adult or sexual output you intend to generate.

At training or upload time you must complete the versioned attestation we present, stating which category applies; the attestation is recorded with a timestamp and the policy version in force. Providing a false attestation is a serious breach of this Policy and may be a criminal offence. Where a dataset is assembled from media generated on Vellux, our stored generation records additionally evidence its synthetic origin; for images from any other source we rely on your attestation and on the enforcement described below.

You may not:

• assemble any dataset or combination of datasets — including the builder's separate face and body inputs — so that the resulting model reproduces the likeness of a real, identifiable person who has not consented (for example by supplying the same real person's photos as more than one input dataset);
• build or upload a model of a public figure, celebrity, or any third party without their verifiable consent;
• produce a minor's likeness, or any sexual output involving a minor's likeness (Section 3);
• produce sexual or intimate output depicting a real person who has not consented to it (Section 4).

The attestation is our baseline control, not a licence. We may at any time require documentary proof of a depicted person's identity, age, and consent, and we may remove, disable, or delete any model, dataset, or Output — and suspend the account — where such proof is not provided or where we have reasonable grounds to doubt an attestation. Any real person whose likeness appears in a character or its Outputs without consent may demand removal at any time via the Complaints process. Optional retained training datasets are held only for as long as the associated character exists and are purged on its deletion (see the Privacy Policy).

7. Contractually prohibited categories (payment-network floor)

Independently of what may be lawful where you are, the following categories are contractually prohibited on Vellux, and this prohibition applies to Outputs and to prompts and search terms alike — including legal, wholly synthetic, and moderated material. These limits are required by the card networks and adult-payment processors whose rules we must satisfy to offer paid Credits and subscriptions (for example Mastercard's Specialty Merchant standard AN 5196 and Visa's VIRP, and the published prohibited-content lists of specialist adult processors). We state their origin honestly: some of these bans are stricter than local law and exist so that Vellux can be accepted by a mainstream adult payment provider.

You may not create, request, upload, generate, or store material depicting or describing:

• incest or sexual activity between close family members;
• bestiality — sexual activity involving animals, including non-human, cartoon, or fantasy creatures;
• rape or any non-consent theme, including depiction of subjects who are sleeping, drugged, intoxicated, hypnotised, unconscious, or otherwise unable to consent, and any content that glorifies or eroticises non-consent;
• human trafficking, sexual exploitation, or coercion;
• scat, bodily waste, or similar excretory content;
• necrophilia, "snuff," or the eroticisation of death;
• extreme violence, mutilation, torture, or gore, and weapons used sexually or to inflict sexual harm;
• hateful or discriminatory content targeting a person or group on the basis of a protected characteristic;
• the likeness of any real, identifiable person without their documented consent, including celebrities and public figures (this restates Section 4 as a payment-network requirement).

We may screen for and block these categories at the prompt and keyword level as well as at the output level, and we reserve the right to refuse, throttle, remove, or disable anything that falls within them. Attempting to circumvent such screening — for example through obfuscated wording, "jailbreak" prompts, coded terminology, or reference uploads chosen to evade it — is itself a breach of this Policy.

8. Other illegal and abusive content

You may not use Vellux to create, request, or store Content or Outputs that:

• facilitate or promote a crime, terrorism, or violent extremism;
• facilitate human trafficking, fraud, forgery, or the production of fraudulent identity or financial documents;
• infringe any intellectual-property or other proprietary right;
• harass, threaten, defame, blackmail, or extort any person;
• are otherwise unlawful under the applicable law of Switzerland, the EU, the UK, or your own jurisdiction.

9. AI-generated content and disclosure

All Outputs are produced by artificial intelligence. Free-tier images carry a visible Vellux watermark; the marking applied to Outputs may evolve as provenance standards and Generator capabilities develop. Technical metadata can be stripped when a file is re-encoded, so the absence of a marking on a file is never evidence that the file is not AI-generated.

Where a watermark or marking is applied to an Output, you must not remove, obscure, crop out, strip, or otherwise defeat it. Marking does not replace your own duties: where the law of a jurisdiction you distribute into requires it — in particular for "deepfake"-type material that appreciably resembles real persons or events — you must disclose the Output as artificially generated or manipulated when you use it outside Vellux. The absence of a visible watermark on a paid Output does not transfer any rights and does not relieve you of this disclosure duty.

10. Enforcement

"Uncensored" governs taste, not legality. To protect the platform, our payment partners, and potential victims, we operate — and reserve the right to operate — automated and manual enforcement despite the permissive stance:

• Screening. We may apply prompt and keyword blocklists, output classifiers, and CSAM hash-matching, and we may throttle or block requests that our systems flag. These operate at the prompt/search level and at the output level.
• Review of flagged material. While we exercise no prior editorial review of the private Gallery, we may access, scan, and review Content and Outputs that are reported, flagged by our systems, or reasonably suspected of breaching this Policy or the law. Using an uncensored tool does not put your Content beyond our reach for legality checks.
• Reporting and action. Anyone — user or non-user — may report a violation through our Complaints, Reports & Takedowns process or at [email protected]. We assess reports, and where we act we provide the affected user a statement of reasons and the redress options described there. Prohibited content is removed or disabled; confirmed illegal or non-consensual material is prioritised, with NCII handled on a 24-hour fast track.
• Suspension and termination. We may remove or disable Content, Outputs, characters, and datasets, and may suspend or terminate accounts, for breach of this Policy — as set out in the Terms of Service. Suspension blocks access and revokes media access while preserving the underlying data.
• Preservation and cooperation. Where required or appropriate we preserve evidence and cooperate with, and respond to lawful requests from, competent authorities. For CSAM and threats to life or safety we report to the relevant authorities as described in Sections 3 and 4 and in the Complaints policy.

We may also refuse service, and reserve the right to remove or disable any Content or Output at our discretion where we reasonably believe this Policy or the law has been breached.

11. Records and 18 U.S.C. §2257 exemption

The Outputs generated on Vellux are synthetic images and video produced by artificial intelligence. Except to the extent a user supplies photographs of a real person, they are not photographs or recordings of actual human beings engaged in actual conduct. Wholly AI-generated Outputs therefore fall outside the definition of matter to which the federal record-keeping requirements of 18 U.S.C. §2257 and §2257A apply, and in no case does Vellux permit any sexual or sexualised depiction of a minor (Section 3). This statement is posted for the avoidance of doubt; nothing in it limits the prohibitions elsewhere in this Policy.

Where you upload photographs of a real, identifiable adult — for example to train a character/LoRA model (Section 6) — you, not Vellux, are the producer of any depiction you then generate, and you are solely responsible for holding that person's consent, for confirming that they are an adult, and for creating and keeping any age-verification records the law of your jurisdiction requires. Vellux does not operate a public showcase of, and does not itself produce for publication, sexually explicit depictions of real persons.

Any records the operator is itself required to maintain in respect of the Service are held by Dennis Müller, c/o Anibarro, Unit #152, Route de Saint-Cergue 24Bis, 1260 Nyon, Switzerland. Questions about this statement may be sent to [email protected]. This statement is also published as a standalone page: 18 U.S.C. §2257 Exemption Statement.

12. Contact

• Reports of prohibited or illegal content, and legal notices: [email protected] (see the full Complaints, Reports & Takedowns process)
• General support: [email protected]
• Data protection: [email protected]

This Policy sits alongside our other legal documents: Terms of Service · Privacy Policy · Refund Policy · Complaints, Reports & Takedowns · 18 U.S.C. §2257 Exemption Statement · Imprint / Legal Notice.