AIGC Acceptable Use Policy (AUP)

1. Scope

This policy covers all prompts, text, reference images or videos submitted to eluai, AI-generated text, images or videos, and their storage, sharing and distribution. It applies together with the Terms of Service to support safe, lawful and responsible AI creation.

2. Content standards and prohibited categories

Generating, editing, uploading, storing, sharing or distributing the following six categories is explicitly prohibited. Do not evade restrictions using indirect descriptions, split prompts, language changes or other methods.

• Pornography / NSFW: pornography, explicit sexual acts, promotion of sexual services, sexual-organ close-ups, sexualized nudity or nudification, including non-consensual intimate imagery, are prohibited.
• Violence / gore: graphic injury, dismemberment, abuse, gore, and incitement, glorification or instructions for violence, self-harm, terrorism, extremism or weapons of mass destruction are prohibited.
• Hate speech: insults, dehumanization, discrimination or violence targeting race, ethnicity, nationality, religion, gender, sexual orientation, disability or similar characteristics, and advocacy of genocide, are prohibited.
• Child-unsafe content (CSAM): zero tolerance for child sexual abuse or exploitation material, sexualization of minors, grooming, sexual contact or exploitation instructions, including fictional, illustrated and AI-generated forms. Synthetic material is not an exception.
• Deepfakes / impersonation: deepfakes of real individuals, including face swaps, fabricated statements or actions and synthetic intimate scenes, are prohibited regardless of consent. Impersonating people, brands, organizations or professionals to deceive others about identity, source or authenticity is prohibited.
• Copyright / trademark infringement: unauthorized copying or adaptation of protected works, misuse of trademarks or brand marks, counterfeit product materials, or generating, removing attribution from or distributing content to infringe intellectual property are prohibited. Authorized creation must still comply with other restrictions.

3. Material sources and conduct

Only upload your own photos or photos of people who explicitly authorized their use, and other materials you are entitled to use. Do not upload another person's photos, voice or private information without consent. Material authorization does not permit prohibited deepfakes or other violations. Do not use the service for scams, harassment, phishing, malware, illegal activity or mass spam.

Retain applicable AI labels or disclose AI generation when publishing. Do not present outputs as authentic photography, real events or human professional work. A model generating content or failing to block it does not mean the platform permits it. Users must still comply with this policy and review outputs.

4. Platform content moderation

Input checks: the platform's application-level content safety blocklist normalizes prompts and matches keywords before requests reach the model, reducing simple evasion using zero-width characters or inserted separators. Matching requests are rejected and recorded in safety audits. The operator maintains and tests the blocklist in the administration console.

Output checks: generated text such as scripts and storyboards is checked against the blocklist. Image and video generation is also subject to the selected upstream model or interface's safety rules. Independent automatic visual moderation covering every image and video is not currently deployed; upstream filtering is not a guarantee of comprehensive review. The operator investigates suspected outputs and reports manually using tasks, materials and audit records, restricting distribution or removing shares when violations are confirmed.

Review and improvement: the operator reviews risks using reports, blocked-request records and abnormal usage, updates rules and tests them. Automated rules may miss violations or flag permitted content; appeals are available below. Review access and record use are limited to investigation, safety and legal needs, under the Privacy Policy.

5. User reporting channel

Report violations, child-safety risks, impersonation or infringement of your rights through support WeChat luyuhao0422, stating 'Content report'. The operator handles this channel manually for reports, safety feedback and appeals. Mark CSAM, terrorism or urgent personal-safety risks as 'Urgent content safety report'.

Provide the task ID or share link, category, time, factual explanation and contact details. Intellectual-property reports should include evidence of rights or authorization. Do not download, copy or forward illegal material such as CSAM; identifying information is sufficient. Do not send passwords, verification codes or unrelated sensitive information.

The operator reviews reports by risk, prioritizing child safety and urgent risks. Accounts or shares may be restricted before further review where needed. Results are communicated where disclosure is permitted, and reporters may be asked for information. Reporter information is limited to investigation and lawful handling, without unnecessary disclosure to the reported party. Malicious or false reports may also result in action.

6. Enforcement actions

Actions depend on severity, repetition, risk and findings. Serious violations may directly result in suspension or termination without graduated warnings.

• Generation blocking: reject requests or stop tasks without delivering blocked content; credits for failed or unused portions are released under task settlement rules.
• Warning and correction: explain minor or first violations and require the content or conduct to stop, be corrected or be deleted.
• Content action: remove violating public content or share links and restrict distribution, retaining necessary investigation and legal evidence.
• Account action: repeated violations or high-risk conduct may result in feature restrictions or suspension. Serious violations such as CSAM, serious deepfakes or fraud may result in permanent termination.
• Legal action: suspected illegal activity or threats to children or personal safety are reported to competent authorities as required by law. Refunds and remaining paid benefits are reconciled under the Terms of Service and law, not automatically forfeited in full.

7. Appeals and policy updates

Appeal blocking, removal, restrictions or termination through support WeChat luyuhao0422, stating 'Content appeal', with your account or task ID, notice, reasons and evidence. The operator re-examines facts and rules and communicates the result, correcting errors. Do not appeal by resubmitting violating content, creating other accounts or evading rules.

Updates reflect service and safety requirements and display the date and version. Material changes are communicated under the Terms of Service, with consent where required. Underlying model names, capabilities and creation options are available on the About page and model catalog. Model rules do not replace this policy.