Acceptable Use Policy

Version 1.0 · 14 August 2026

Status: working document. This policy describes what the platform actually enforces today and is written to be operationally true rather than decorative. It has not been reviewed by a qualified lawyer. Obtain that review before relying on it with any external customer, and before publishing it as a binding term.

OnlyHub produces synthetic and synthetically-altered media of people. That capability is legitimate and it is also the capability behind some of the most serious harms online. This policy sets the line, and much of it is enforced by the software rather than left to trust.

1. Absolutely prohibited

These are not judgement calls. Accounts are terminated and, where the law requires it, reported.

Any sexual content involving a minor, real or synthetic, including any depiction of an adult styled to appear to be a minor.
Intimate imagery of a real person without their documented consent (non-consensual intimate imagery). Creating this is a criminal offence in the United Kingdom and in many other jurisdictions.
Applying a real person's likeness without a signed release held on file and referenced in the consent record.
Impersonating a real, identifiable individual, or presenting synthetic media as authentic footage of them.
Content that incites violence or hatred, or that depicts non-consensual acts, or that is unlawful where it is produced, hosted or published.

2. Not permitted without explicit arrangement

Uploading source footage you do not own or have a licence to process.
Publishing to any platform in breach of that platform's own terms, including its rules on AI-generated content and account ownership.
Removing, altering or suppressing the provenance manifest attached to an output.
Disabling, bypassing or working around the consent gate, the approval requirement, or the pre-publish compliance check.

3. What the platform enforces for you

These are not promises of good behaviour; they are refusals built into the system, and they are covered by automated tests.

Control	Behaviour
Consent gate	A persona cannot be approved without an active consent record, reference material and a canonical embedding.
Re-check at execution	Consent and approval are verified again when work runs, not only when it is requested — so a revocation stops queued work.
Revocation cascade	Withdrawing consent retires every persona version it authorised.
Human approval	AI output cannot be marked publishable without a person approving it.
Disclosure check	Synthetic content destined for an account that requires an AI label is refused if the label is missing.
Authenticity claims	Captions asserting content is “100% real”, “not AI” or similar are refused on synthetic media.
Licence gate	A generation model whose own licence forbids this use case cannot be selected.

4. Age

OnlyHub is for adults. Every person depicted, and every person operating an account, must be 18 or over. Where output is distributed to the public, the operator is responsible for age assurance meeting the requirements of the destination jurisdiction — in the United Kingdom this includes duties under the Online Safety Act.

5. Reporting and enforcement

Suspected misuse should be reported to the operator immediately. Violations of section 1 result in immediate suspension, preservation of the audit record, and referral to law enforcement where applicable. The append-only activity log exists precisely so that questions of who did what can be answered.