Acceptable Use Policy & AIGC Guidelines

1. Scope and relationship to the Terms

This policy applies to every upload, prompt, generation request, and any distribution or sharing of content through Gemini Omni Editor, including attempts to bypass safety controls. Violations of this AUP are violations of the Terms.

Current enforcement relies on an automated server-side policy gate plus the AI provider's built-in safety filters. We do not claim 24/7 human moderation of every generation. Logs and provider signals may be used to review severe or repeated violations; see Section 5.

2. Prohibited content — six classes

You may not upload, generate, request, or distribute content in the following classes, regardless of whether it is text, image, or video. The same prohibitions apply to prompting, uploading source material, or sharing outputs:

• Pornographic / NSFW: sexual content intended to be explicit, including graphic sexual acts, explicit nudity, or sexual solicitation involving real or synthetic subjects;
• Violence / gore: graphic cruelty, gore, instructions or encouragement for violent wrongdoing, or detailed depiction of serious physical harm;
• Hate speech: content that demeans, dehumanizes, or incites hatred or discrimination against protected characteristics such as race, ethnicity, religion, gender, sexual orientation, disability, or nationality;
• Child safety / CSAM: any sexualization, grooming, sexual solicitation, or sexual depiction of minors, and any generation, storage, or distribution of child sexual abuse material (CSAM) — zero tolerance;
• Deepfake / impersonation: synthetic portrayal of a real person without that person's consent, especially where it could mislead, defraud, harass, or harm reputation;
• Copyright / trademark / IP: material that infringes another's copyright, trademark, or other intellectual property, including unauthorized reproduction, distribution, or removal of rights management information;
• These prohibitions cover uploads, prompts, generation requests, and distribution. Attempting to evade them by rephrasing, encoding, splitting, or combining inputs remains prohibited. See Section 3.

3. Bypass and evasion is prohibited

Do not attempt to circumvent safety and access controls, including:

• prompt injection, jailbreaks, or instructions that tell the model to ignore or override its safety rules (e.g., 'ignore previous instructions', 'jailbreak', 'DAN mode', 'unfiltered mode');
• encoding, splitting, or obfuscating prohibited requests to evade filters;
• uploading prohibited material via an innocuous prompt, or sharing prohibited outputs outside the service.
• Bypass attempts may be blocked by the same automated gate and are themselves policy violations.

4. Enforcement

When a request or content is identified as violating this AUP, we may:

• refuse to generate or process the request — blocked requests are not charged and no credits are deducted;
• remove, block, or disable access to violating uploads or outputs;
• restrict features (for example, temporary limits on generation) after repeated violations;
• suspend or terminate accounts engaged in serious or repeated violations;
• retain minimal security logs needed to review the event — category and timestamp, without storing the full prompt or unnecessary private data — and handle them under the Privacy Policy;
• cooperate with law enforcement or rights holders where legally required or permitted, for example to comply with valid legal process or to address imminent harm or valid IP claims;
• For severe illegal content (for example, CSAM or instructions enabling serious violence), we may act immediately without prior notice and refer the matter as required by law.

5. Automated detection and provider safety

The service applies (a) a shared server-side automated policy gate that screens text prompts before credits are deducted, and (b) the AI provider's own safety filters that may block generation or return a safety refusal during processing.

When the provider refuses on safety grounds after credits were deducted, the credit-compensation path automatically refunds the credits for that generation. A blocked request from the pre-generation gate never deducts credits in the first place.

Logs for policy blocks do not retain the full prompt or sensitive personal data; they record only a desensitized fingerprint, category, length, and timestamp for abuse-prevention.

6. Reporting and appeal

If you believe content violates this AUP, or if your request was blocked in error, you can report or appeal through:

• in-app Contact page at /contact;
• email: support@geminiomnieditor.com.
• We review reports and appeals using the available account, generation, and security records. Submit an appeal promptly and include enough detail for us to locate the request.
• You may also use these channels to report impersonation, copyright/trademark concerns, or other rights-based issues.

7. What to include — and what not to include

To help us act quickly, include:

• the URL, generation ID, or task ID of the content (if any), and the approximate time of the request;
• a brief, factual description of why the content may violate this AUP;
• your contact email and, for IP claims, proof of ownership or authorization.
• Please avoid sending unnecessary sensitive data. Redact full card numbers, passwords, private keys, government IDs, or other secrets. Do not include the full original prompt if it contains private information — summarize the concern instead. For child-safety or IP reports, attach only what is minimally necessary to identify the issue.

8. Emergency and child-safety matters

If a situation involves a real child at risk, CSAM involving a real minor, or an imminent threat of serious real-world harm, contact your local law enforcement immediately in addition to reporting to us.

Do not attempt to collect, store, or redistribute suspected CSAM to 'document' the violation. Instead, report the reference (URL, generation ID, timestamp) and cease further handling.

9. Changes and contact

We may update this AUP as the service, provider filters, or legal requirements evolve. The version posted on /aup is the current version. For questions about this AUP, contact us at support@geminiomnieditor.com or via /contact.