ContentPolicy

When using this Service, you must not: via prompts, uploaded images, or other means, attempt to generate, edit, enhance, or distribute the following content:

• NSFW / Adult content: Nudity, sexual acts, sexually suggestive poses, or depiction of sexual organs;
• Minors: Any sexualized, inappropriate, or suggestive content involving minors ("zero tolerance");
• Sexualized real persons: Sexualized edits, face swaps, or "undressing" type operations on real persons (including celebrities and private individuals);
• Violence and gore: Extreme violence, torture/murder, self-harm, gory accidents, and other disturbing imagery;
• Hate and discrimination: Inciting hatred or discrimination against groups based on race, religion, gender, sexual orientation, disability, etc.;
• Terrorism and extremism: Terrorist organization logos, propaganda, and inciting symbols;
• Illegal activities: Depiction of illegal activities such as drugs, weapons manufacturing, fraud, or money laundering;
• False and deceptive: Deepfakes used for fraud, fake identification documents, or fabricated news footage;
• Infringing content: Content that infringes on others' copyright, trademark, portrait, or privacy rights;
• Other content prohibited by the laws of China or your jurisdiction, or by the jurisdiction where this Service operates.

To enforce the above prohibitions, all prompt-based generation endpoints of this Service (including but not limited to AI image editing, prompt optimization, etc.) are integrated with Creem Content Moderation API for review before calling the AI model:

• allow: passes moderation and enters the normal generation flow;
• deny: clearly violating, immediately blocked, model not called;
• flag: in the monitored gray zone, per Creem's official recommendation also blocked, model not called;
• Fail closed: When the Moderation service itself is unavailable (network failure, 5xx, timeout, etc.), we also reject generation rather than allowing it through, to prevent any violating content from slipping through.

Moderation calls have a 5-second timeout, and all moderation decisions carry a unique ID for audit traceability.

When the system detects a violation attempt, we will take the following measures based on severity:

• First violation: Reject the current generation and return a moderation-failed notice to the user;
• Repeated violations: Restrict account features, deduct violating credits, or suspend the account;
• Severe violations (e.g., involving minors, terrorism, illegal activities): immediately and permanently ban the account, and report relevant information to law enforcement authorities in accordance with the law;
• Fund handling: Severely violating accounts will not be refunded for unused credits, and paid amounts may be frozen for dispute compensation.

• You must ensure your prompts and uploaded images do not contain prohibited content; technical moderation is only a supplementary measure;
• You must not attempt to bypass moderation through spelling variants, coded language, emojis, foreign languages, indirect descriptions, or other means;
• You must not perform any AI editing on images of minors;
• If you discover others' violating behavior or generated violating content, please report via [email protected].

If you believe your prompt was incorrectly blocked, you can submit an appeal via [email protected], including:

• Account email and user ID;
• The time of blocking and the prompt content;
• Your reasoning for why it should be allowed.

We will review and respond within 3 business days. During the review period, the relevant features may remain restricted.

This policy may be updated from time to time based on Creem platform requirements, legal or regulatory changes, or risk considerations. Material changes will be announced in a prominent place within the Service. Continued use of the Service constitutes acceptance of the updated policy.

For reporting and appealing content safety issues:
Email: [email protected]