Back to Legal

03 - Acceptable Use Policy

Last updated: 2026-09-23

This Acceptable Use Policy ("AUP") sets the rules for what you may create, upload, store, generate and share using PyxelJam.Studio ("PXLJ"). It forms part of the Terms of Service, and breaching it is a material breach of those Terms.

PXLJ is a filmmaking tool. Like every powerful creative tool it can be misused, and generative video can be misused in ways that hurt real people badly. These rules exist to protect those people, our other customers, our model providers and the Service itself.

Dramatic content is not the target of this policy. Films are about crime, war, addiction, cruelty and death. Depicting a difficult subject in a narrative is not a breach. Using the tool to actually harm a real person is.

If something you want to make sits in a grey area, ask us first at support@pyxeljam.studio. Asking is always cheaper than a termination.


1. Scope

This policy applies to everything in your account: prompts, scripts, uploaded images, audio and video, recorded performances, character and location library entries, generated Outputs, exports, share links, project names and support messages. It applies to you, to every seat on your account, and to anyone you give a share link to.

2. Illegal content and conduct

Do not use PXLJ to create, upload, store, generate or distribute content that is illegal under United States federal or applicable state law, under the law where you are, or under the law where your audience is. This includes, without limitation:

  • content that facilitates serious crime, including instructions for weapons, explosives or attacks intended to cause harm;
  • content that breaches export control, sanctions or counter-terrorism law;
  • terrorist or violent extremist content, or material glorifying or recruiting for it;
  • content that infringes another party's intellectual property (see Section 9);
  • fraud, phishing, malware distribution, or content designed to deceive people into parting with money or credentials.

3. Child safety - zero tolerance

Any content that sexualises a minor is absolutely prohibited. This covers photorealistic generated imagery, stylised or animated depictions, written material, audio, and any attempt to generate it whether or not the attempt succeeded. It covers real children and entirely synthetic depictions alike.

It also covers:

  • uploading images of real children as reference or character material for sexualised generation;
  • prompts, character descriptions or age settings that attempt to produce a sexualised minor;
  • attempts to circumvent our controls or a provider's safety filter to reach such material.

There are no exceptions, no artistic defence, and no warnings.

Consequences. Immediate and permanent account termination without refund; preservation of the account and its content as evidence; and reporting to the National Center for Missing and Exploited Children (NCMEC) through the CyberTipline, and to any other competent authority. As a US provider we are required by federal law (18 U.S.C. section 2258A) to make that report, and to preserve the material for the statutory period. We will cooperate fully with law enforcement.

4. Real people - likeness, voice and deepfakes

This section is about identifiable real people, living or recently deceased.

4.1 Consent is required

Do not upload, record, or generate content depicting a real, identifiable person unless you hold their informed consent for that use, or you have another lawful basis. This applies to their face, their body and their voice, and it applies equally to reference photographs, character library entries, and recordings made with the Acting feature. See the AI Content Terms, Section 7.

4.2 Voice cloning

Do not create a synthetic voice from a real person's speech without that person's explicit consent. Do not use a cloned voice to say things the person did not say and would not agree to say.

4.3 Deepfakes and impersonation

Do not generate content depicting a real, identifiable person and present it, or allow it to be presented, as genuine. This includes fabricated speeches, statements, endorsements, confessions, phone calls, interviews and events.

Narrow exception. Clearly labelled satire, parody, commentary or dramatic portrayal with a lawful purpose is permitted, provided a reasonable viewer would understand the content is not genuine, and provided the labelling stays attached when the content is shared. You are responsible for keeping that context intact. The exception does not apply to Section 4.4 or Section 4.5.

4.4 Intimate and sexualised depictions of real people

  • Do not create or upload intimate or sexual imagery of any real person without their explicit, documented consent.
  • Do not sexualise an identifiable real person, including public figures, in generated content. "It is obviously fake" is not a defence; the harm to the person is real. In the United States the federal TAKE IT DOWN Act criminalises the publication of non-consensual intimate imagery including AI-generated depictions and requires platforms to remove it on notice, most US states have their own statutes, and the same conduct is criminal in most of our customers' countries.
  • Do not use PXLJ to undress, body-swap, or otherwise manufacture intimate depictions of anyone.

4.5 Political and electoral deception

Do not generate synthetic media of candidates, officials or election processes intended or likely to mislead voters. Do not fabricate footage of an official announcement, a concession, a scandal or a public safety event.

4.6 Deceased people

Recreating a deceased person's likeness or voice requires the consent of their estate or their family, and is unlawful in some jurisdictions regardless.

4.7 Identification, surveillance and profiling

Do not use PXLJ, or anything you make with it, to identify, locate, track or build a profile of a real person from their face, body or voice. This includes matching an uploaded likeness against other material to work out who someone is, inferring characteristics such as health, disability, race, religion, political opinion, sexual orientation or immigration status from an image or a recording, and passing any of that to a third party. It applies whether or not the person consented to appear in your production.

A reference photograph you are entitled to use is there to make a film with. It is not a search key.

5. Adult and sexual content

Sexually explicit content is not permitted on PXLJ at this time. Do not generate, upload or store pornographic or sexually explicit material, whether photorealistic, stylised or animated, and whether the subjects are real or entirely synthetic.

What is permitted: nudity and sexuality handled as narrative film handles them - implication, tasteful framing, romance, intimacy that stops short of explicit depiction, non-sexual artistic or medical nudity.

What is not: explicit sexual acts, pornographic framing, or content whose primary purpose is sexual arousal.

Why this rule exists, plainly. The models we route to enforce their own content policies, our payment processor's rules restrict adult content, and we have not built the age verification, geographic restriction, model routing and moderation infrastructure that hosting adult content responsibly requires. Permitting it today would be a promise the product cannot keep.

6. Harassment, hate and violence against real people

Do not use PXLJ to:

  • harass, bully, threaten, stalk or intimidate anyone;
  • create content that dehumanises or incites hatred or violence against a person or group based on race, ethnicity, national origin, religion, disability, gender, gender identity, sexual orientation, age or immigration status;
  • doxx anyone, or generate content that reveals private information about a person without their consent;
  • depict real, identifiable people being harmed, killed, or subjected to degrading treatment;
  • promote or encourage suicide, self-harm or disordered eating, give instructions or methods for any of them, or aim that material at a person or an audience likely to act on it.

Depicting hatred, violence, bigotry or self-destruction within a narrative, as subject matter, is not a breach: films have always gone to those places, and a story about a character's despair is not encouragement. Producing it at a real person, or as a message to a real audience, is.

7. Disinformation and synthetic media presented as real

Do not present generated content as authentic footage of real events. In particular, do not fabricate:

  • news footage, breaking-news reports or journalistic material;
  • public safety emergencies, disasters, attacks or crimes;
  • statements or footage attributed to real organisations, governments, companies or media outlets;
  • scientific, medical or financial claims dressed as documentary evidence;
  • reviews, testimonials or endorsements from people who did not give them;
  • crowds that do not exist: sock-puppet personas, invented spokespeople, manufactured grassroots support, or coordinated posting campaigns that use generated faces and voices to make one opinion look like many.

Fiction is fine. Fiction wearing the costume of the news is not.

You are separately responsible for complying with disclosure rules that apply to your work, including the EU AI Act's transparency obligations for synthetic media and the AI-disclosure requirements of the platforms you publish on.

8. Platform integrity

Do not:

  • attempt to circumvent credit metering, tier gates, project limits, storage caps, rate limits or budget caps;
  • share one seat between multiple people, or resell access to your account;
  • resell PXLJ generation capacity as a service of your own, or wrap the Service in your own product, without a written agreement with us;
  • create accounts by automated means, or create multiple accounts to obtain repeated trial credits;
  • scrape, crawl, reverse engineer, or attempt to extract our prompts, model routing logic or system instructions;
  • use the Service, or anything it generates, to develop, train, fine-tune, evaluate or benchmark a competing generative model, or extract Outputs in bulk to assemble a training or evaluation dataset;
  • probe, scan or test the security of the Service without our written permission, or attempt to access another account's data;
  • interfere with the Service, overload it, or use it to attack anything else;
  • deliberately craft prompts to circumvent a model provider's safety filter in order to reach content this policy prohibits.

9. Intellectual property

Do not:

  • upload material you do not have the rights to use;
  • deliberately generate content that infringes a third party's copyright, trademark or design rights;
  • prompt for a copyrighted character, a protected work or a specific artist's signature style with the intent of passing the result off as, or substituting for, the original;
  • use PXLJ to produce counterfeit goods, fake branding or forged documents;
  • remove or falsify attribution, provenance or watermarking that a rights holder attached to material you use;
  • defeat digital rights management or any other technical protection on material you want to bring in, or upload material you obtained that way.

Style is not ownable, but a character is. Making a film that looks like a 1970s Italian thriller is fine. Generating a specific studio's copyrighted character is not.

9.1 Copyright notices (DMCA)

If you believe content on PXLJ infringes your copyright, send a notice to our designated copyright agent at support@pyxeljam.studio, with "DMCA notice" in the subject line, containing the elements required by 17 U.S.C. section 512(c)(3): your signature (electronic is fine), identification of the copyrighted work, identification of the material you want removed and where it is, your contact details, a statement that you have a good faith belief the use is not authorised, and a statement under penalty of perjury that the notice is accurate and that you are the rights holder or authorised to act for them.

We will remove or disable access to material on a valid notice, tell the person who posted it, and restore it if they send a valid counter-notice and the complainant does not file suit. We terminate the accounts of repeat infringers.

For anything that is not a copyright claim - trademark, publicity rights, a likeness used without consent - write to support@pyxeljam.studio with enough detail to identify the material and your rights in it.

If you are in the EU, the same address is our notice-and-action channel under the Digital Services Act for any content you believe is illegal, not only copyright. We acknowledge every notice when we receive it. Whenever we remove content or restrict an account, whether because of a notice or on our own initiative, we give the person affected a statement of reasons: what we did, which rule it relates to, what content or conduct it concerned, and how to appeal (see Sections 11.3 and 12).

10. Content we may block regardless

Model providers run their own safety filters and may refuse a generation that this policy permits. We cannot override those refusals. See the AI Content Terms, Section 10.

11. Enforcement

11.1 How we find out

We do not proactively monitor or review your projects. We act on reports from people affected, on notices from model providers, on automated abuse signals, and on legal requests.

11.2 What we may do

Depending on severity, whether it is a first occurrence, and whether it looks deliberate, we may:

  1. Warn - tell you what breached the policy and ask you to remove it;
  2. Remove - delete the specific content or revoke a share link;
  3. Restrict - disable a feature, a model, or your ability to share publicly;
  4. Suspend - pause the account while we investigate;
  5. Terminate - close the account permanently, without refund;
  6. Report - notify law enforcement.

For most breaches we will warn first. For Section 3 (child safety) there is no warning: termination and reporting are immediate. For serious breaches of Sections 4.4, 4.5 and 6 we may terminate without a prior warning.

11.3 What you will be told

Unless the law prevents us from doing so, or telling you would prejudice an investigation, we will tell you what action we took, which section of this policy it relates to, what content or conduct it concerned, and how to appeal.

11.4 Credits and refunds on enforcement

Termination for a breach of this policy does not entitle you to a refund of subscription fees, and trial and subscription credits are forfeited. Credits you bought in a credit pack and have not consumed are forfeited only when the termination is for a breach of Section 3 (child safety) or a serious breach of Section 4 (a real person's likeness or voice used without consent, a sexualised depiction of a real person, or political deception). For a termination on any other ground we refund the unconsumed balance of purchased credit packs to the original payment method.

12. Appeals

If you believe we got it wrong:

  1. Email support@pyxeljam.studio within 30 days of the action, from the account's registered email address.
  2. Tell us what was actioned and why you think it was wrong. Include context we would not otherwise have: the consent you hold, the licence you hold, the creative context, the labelling.
  3. We will acknowledge within 5 business days and decide within 30 days.
  4. A person reviews every appeal. Appeals are not decided by automated means.
  5. If we were wrong, we will reinstate the account or the content and re-credit anything wrongly consumed.

Appeals against a Section 3 termination will be reviewed, but the account remains closed and the report to authorities stands while the review runs.

If you are still unhappy after an appeal, you may take the matter to a court in the venue named in the Terms of Service, Section 16. If you are in the EU, you may also use an out-of-court dispute settlement body certified under the Digital Services Act.

13. Reporting a violation

To report content or conduct that breaches this policy, email support@pyxeljam.studio with the share link or account concerned, what you believe is wrong, and how to reach you.

To report suspected child sexual abuse material, contact us and report it directly to the NCMEC CyberTipline (https://report.cybertip.org, or 1-800-843-5678) or to the authority in your country. Do not download, copy or forward the material.

14. Changes

We may update this policy as the law, the models and the misuse patterns change. Material changes are notified as described in the Terms of Service, Section 15.


Report a violation or ask a question: support@pyxeljam.studio