Content Policy and Moderation

Lumen Weaver generates Minecraft-style character skins from text and reference images. This page explains what you may and may not create, what we do to keep the Service safe, and how to report a problem. It supplements Section 6 of our Terms of Service. The Service is operated by Lumen Interactive Limited, incorporated in Hong Kong (Company Registration No. 81032006).

1. What the Service produces

Output is limited to 64×64 or 128×128 pixel character-skin sheets in the Minecraft layout. The models are trained on stylised, blocky character art; they do not produce photorealistic images. Every result is generated by an AI system without human review before delivery, and the same input can produce different results.

2. Prohibited content and use

You may not use the Service to create, upload or attempt to create:

• Sexual content involving minors, or any content that sexualises, exploits or endangers children. This includes stylised or "anime" depictions. Zero tolerance: accounts are terminated and, where the law requires, reported.
• Sexually explicit content or nudity.
• Hateful content — symbols, slogans or depictions that promote hatred or violence against people based on race, ethnicity, religion, nationality, disability, sex, sexual orientation or gender identity.
• Violent extremism or terrorism, including insignia of proscribed organisations.
• Harassment or defamation of a real person.
• Likenesses without consent — reference images of a real, identifiable person who has not agreed to their likeness being used (for a minor, whose parent or guardian has not agreed).
• Infringing content — copyrighted characters, logos or designs you do not hold the rights to, when the result would infringe the rights holder's rights; trademarks used in a way that suggests endorsement.
• Deceptive use — impersonating a real person or organisation, or presenting generated output as an authentic photograph or record.
• Abuse of the Service — bulk or automated generation to build datasets, training other AI systems on the output, operating multiple accounts to evade limits, or attempting to extract model weights or source code.

3. Reference images and real people

The Service does not perform facial recognition and does not try to identify people. Reference images are used only to pick up colours, silhouette and style for a blocky skin. Uploads are retained for a limited period (see the Privacy Policy) so that reports can be investigated.

4. How we enforce this policy

We are a small team and we are transparent about what our enforcement actually consists of today:

• Mandatory confirmation before Image-to-Skin (Section 3), recorded with each request.
• Rate limits and abuse detection on the server: per-account and per-network limits on generations and uploads, detection of repeated identical prompts or repeatedly re-uploaded images, and blocking of disposable email domains at sign-up. These protect against bulk abuse and dataset farming.
• Prompt length and format limits; prompts are parsed by a language model into a structured style description before generation, which discards free-form content that does not describe a character's appearance.
• Audit trail: prompts, reference images, generated results and the account that requested them are retained for the periods stated in the Privacy Policy, so that every reported item can be traced to a request and an account.
• Human review of every report (Section 5).
• Account action: warnings, removal of content from history, suspension, or termination, in proportion to the severity of the breach. Child-safety and non-consensual-likeness violations lead to immediate termination.

We do not currently run an automated image classifier on uploads or outputs. We are evaluating automated screening for uploaded images and will update this page when it is in place. Until then, enforcement relies on the measures above and on reports from users and rights holders.

5. Reporting content

To report a skin, a reference image, an account or a prompt, email [email protected] with:

• a link, screenshot or description of the content and, if you have it, the generation date or task ID;
• which rule in Section 2 you believe it breaks;
• for likeness or intellectual-property complaints, a statement that you are the person shown or the rights holder (or authorised to act for them) and how we can contact you.

What happens next:

1. We acknowledge your report within 2 business days.
2. A member of our team reviews the content against this policy. For child-safety reports the content is removed immediately on receipt, before review is complete.
3. We decide within 7 days for most reports and tell you the outcome. Where we remove content or act against an account, we also tell the account holder what was removed and why, unless the law prevents this.
4. If you disagree with the outcome, reply to our decision and a different team member will re-review it.

Rights holders may also send takedown notices for copyright or trademark infringement to the same address. We honour valid notices and may terminate accounts of repeat infringers.

6. Law enforcement and legal requests

We respond to valid legal requests from competent authorities in accordance with Hong Kong law and the law applicable to the requesting authority, and we notify affected users where the law permits.

7. Changes

We will update this page as our moderation tooling evolves. The "Last updated" date shows the current version.

8. Contact

Lumen Interactive Limited (Hong Kong, No. 81032006)
Unit 1610, 16/F, Tower 2, Silvercord, 30 Canton Road, Tsim Sha Tsui, Kowloon, Hong Kong
Email: [email protected]