Acceptable Use Policy

This policy applies to everything that passes through Breme: prompts you write, media you upload, content you generate or edit, and anything you publish or share from the service. It applies to every plan and every feature.

It is incorporated into our Terms of Service at breme.ai/legal/terms. By using Breme you agree to both.

If you violate this policy, we may remove content, limit features, or terminate your account, depending on severity. Where the law requires it, we report illegal content to the appropriate authorities.

We have zero tolerance for child sexual abuse material and for any sexualization of minors. This includes generated, edited, fictional, stylized, or cartoon depictions. There are no exceptions and no warnings.

We report this material and the accounts behind it to the National Center for Missing and Exploited Children (NCMEC) and to law enforcement, and we terminate the account immediately.

You may not create or share sexually explicit content depicting a real, identifiable person, or non-consensual intimate imagery of any kind. Consent from the person depicted does not make explicit content of real people acceptable on Breme; we do not host it.

This includes undressing tools, face swaps onto explicit material, and any content that sexualizes an identifiable person without their involvement.

Breme does not generate, edit, or host sexually explicit content. That covers exposed genitalia, sexual acts, erotica, fetish material, and content made to arouse, whether the people in it are real, fictional, stylized, or animated. There is no adult mode, no age-verified tier, and no private exception. The same standard applies in every plan and every model.

The line is the one film has always drawn. Think of the standard as PG-13 and R: romance, kissing, swimwear, lingerie in a fashion context, period or artistic nudity implied by framing, and the blood, injury, and horror that legitimate drama calls for are allowed. Content that would rate NC-17 or be sold as pornography is not.

Mature material that is allowed, such as graphic film violence or implied nudity, may be labeled Mature. Mature content stays private to your account, is never featured on our public pages, and is never used to improve our models.

Every model we run screens prompts, reference media, and results against its provider’s content filters, and we keep those filters on their strict settings. A refusal is recorded on your account. Repeated attempts to produce explicit content lead to a cooling-off period and then to closure of the account. Trying to get around a refusal by rewording, re-uploading, or switching models counts as an attempt.

Do not use Breme to:

• Promote, glorify, or produce propaganda for terrorism or violent extremism.
• Incite, threaten, or celebrate violence against any person or group.
• Harass, bully, or intimidate anyone, including through repeated unwanted depictions of them.
• Demean, dehumanize, or promote hatred against people based on race, ethnicity, religion, nationality, disability, sex, gender identity, sexual orientation, age, or any other protected characteristic.

Depicting violence for legitimate creative work, such as film, documentary, or commentary, is allowed within reason. Content whose purpose is to threaten, recruit, or degrade real people is not.

Do not use Breme to defraud or deceive people. This includes:

• Scams, phishing, and social engineering, including fake voices or videos used to trick someone into sending money or revealing information.
• Impersonating a person, company, or institution in a way intended to deceive.
• Fake reviews, forged documents, counterfeit credentials, or fabricated evidence.

Do not present fabricated content as fact. This includes fabricated depictions of elections, election processes, emergencies, disasters, or public figures that are intended to deceive viewers about real events.

Fiction, satire, and speculative work are welcome. The line is intent and presentation: if content is designed to make people believe something false actually happened, it does not belong on Breme.

Do not create, upload, or share content that infringes copyright, trademark, or publicity rights. That includes uploading media you do not have rights to, generating content built around protected characters or marks for commercial use, and trading on someone’s name or likeness without permission.

If you believe content on Breme infringes your rights, our copyright process is described in the Terms of Service, and you can reach us at info@breme.ai.

Do not use Breme to create or distribute malware, or to produce instructions for serious harm, such as the synthesis of weapons, explosives, or dangerous chemical, biological, or radiological agents.

Do not abuse the platform itself:

• No spam, bulk unsolicited content, or engagement manipulation.
• No scraping the service or harvesting other users’ content or data.
• No reverse engineering the service or probing it for vulnerabilities outside a sanctioned disclosure.
• No circumventing usage limits, safety filters, or content moderation, whether through prompt manipulation, multiple accounts, or technical means.

Depicting a real, identifiable person in generated content requires their consent. For public figures, clear satire or commentary is allowed when the context makes it obvious that no reasonable viewer would take the content as real.

Editing a photo or video of a real person into revealing, intimate, or sexual content is never allowed, with or without their consent, and models that detect a real face in a reference may refuse such an edit outright.

Cloning or imitating someone’s voice requires that speaker’s permission. This applies to private individuals and public figures alike. Before Breme clones a voice from a recording you supply, you confirm that the voice is your own or that you hold the speaker’s written permission. Cloning the voice of a public figure or a performer you do not represent is not allowed even with a recording in hand.

Realistic AI depictions of real people, places, or events must be disclosed as AI-generated where the law requires it, including under the EU AI Act, and must never be presented as authentic footage, audio, or photography.

Breme may embed machine-readable metadata in exports marking them as AI-generated. This helps platforms, journalists, and viewers tell synthetic media from authentic recordings.

Removing, altering, or tampering with this metadata to pass content off as authentic is prohibited.

Breme builds on models from OpenAI, Anthropic, Replicate, and the model vendors Replicate hosts, along with Recraft, ElevenLabs, and Fish Audio. Your use of Breme must also comply with those providers’ usage policies, which we are contractually required to pass through to you.

We may block prompts or content that a provider prohibits, even where this policy does not spell out the specific case.

Every prompt, reference, and result passes through the content filters of the model that produces it, pinned to their strict settings, and refusals are recorded against the account. Reports are reviewed by a person. Enforcement is proportionate: a first accidental violation is handled differently from deliberate, repeated abuse, and anything involving minors or intimate imagery of a real person ends the account on the first occurrence.

To report content you believe violates this policy, email info@breme.ai with a link or description of the content. Intimate imagery of a real person, including sexual deepfakes, has its own fast path: see breme.ai/legal/take-it-down. We act on valid requests within 48 hours.

If we take action against your content or account and you believe we got it wrong, you can appeal by emailing info@breme.ai; a human will review the decision.

We may update this policy as the law and our model providers’ policies evolve. The effective date at the top reflects the latest version, and continued use of Breme after an update means you accept the revised policy.