Content Policy
This policy sets out what you can and cannot make with Poze. It forms part of the Terms of Service. Moderation checks three points: what you upload, what you write in a prompt, and what comes out the other end.
1. Absolutely prohibited
These result in immediate termination. Where a criminal offence may be involved we preserve evidence and report it to the appropriate authority, without notifying the account holder.
- Child sexual abuse material, or any sexualisation of a minor — including making adult material appear to depict a minor, or ageing a real minor's image forward.
- Non-consensual intimate imagery, including placing someone's face onto sexual content.
- Terrorist or violent extremist material, or content that incites violence against a person or group.
- Content created to commit a crime — impersonating someone to obtain money, forging identity or financial documents, or facilitating human trafficking.
2. Faces and voices need permission
This is where a creative AI app most often causes real harm, so it gets its own section.
- Before using someone else's face for an avatar, face swap, motion transfer or template, you need their clear permission.
- Voice cloning is for your own voice only. You confirm this when you submit, and we keep that confirmation.
- Public figures are off limits — celebrities, politicians, executives, influencers, anyone recognisable. Moderation blocks them, and the app stores require this of us.
- Do not create content that could lead a reasonable person to believe someone really said or did something. Satire and fan work are fine when they are obviously creations.
- Deceased people still have families, and many places grant post-mortem personality rights. Treat them as you would the living.
3. Restricted content
These will not necessarily get you banned, but they are blocked at generation, kept out of the community feed, or both.
| Category | Where the line sits |
|---|---|
| Sexual content and nudity | Not permitted, generated or published. Artistic or medical framing does not change this. |
| Violence and gore | Realistic injury, blood, torture or cruelty to animals is not permitted. Stylised action is fine. |
| Hate and harassment | No attacks based on race, ethnicity, national origin, religion, caste, sexual orientation, sex, gender identity, disability or serious disease. No targeting, humiliating, doxxing or threatening a specific person. |
| Self-harm | Nothing that encourages or glorifies suicide, self-harm or disordered eating, and no depictions of methods. |
| Misleading content | No fabricated news footage, election disinformation, or content presented as an authentic photograph or recording. |
| Dangerous advice | No medical, legal, financial or safety instructions presented as authoritative. |
| Regulated goods | No promotion of drugs, weapons, explosives, illegal gambling or forged documents. |
| Political campaigning | Poze may not be used to produce campaign material or state propaganda. |
| Intellectual property | Do not reproduce protected characters, logos or artwork in a way that infringes, and do not imitate a living artist's style in order to pass work off as theirs. |
4. Erasing, watermarks and AI marking
Remove text and AI erase exist to tidy unwanted text, bystanders and clutter out of your own material.
- Do not use them to strip a watermark, credit or copyright notice from someone else's work. In most jurisdictions that is infringement, and in some it is a separate offence.
- Do not use them to remove the AI-generated marking applied by Poze or any other tool.
- A Pro subscription removes the visible badge from your own Outputs. The embedded watermark and provenance metadata stay — that is a legal requirement, not a paywall.
5. How this is enforced
- Automated review at three points — uploads, prompts and results. A hit blocks the job. You are not charged for a blocked prompt, and a job blocked at the output stage is refunded.
- User reports. Every post and every profile can be reported, and you can block another user so their content stops reaching you. Reports go to a human review queue.
- Escalation, in order: blocking one job; removing the content; limiting features; suspending the account; permanent termination. Violations of §1 skip straight to the end.
- Repeat violations count. A pattern of borderline content is treated more severely than one mistake.
6. If we got it wrong
Automated moderation misfires, and we would rather hear about it. Email support@yes366.com with "Moderation appeal" in the subject line, the UID from the bottom of your Settings screen, and a link or screenshot of what was blocked. A person reviews it and replies.
If a block was wrong, we restore access and refund anything that was charged. If it was right, we will tell you which rule applied so the next attempt does not hit the same wall.
There is no appeal for §1 violations.