Acceptable Use Policy

1. Prohibited content and conduct

You must not use the service, or assist others in using the service, to generate, distribute, or facilitate any of the following. Each category is explicitly prohibited:

1.1 Sexually explicit and NSFW content

Pornography, erotic or sexually suggestive content, sexual services, and any sexual content involving minors or persons who cannot consent. This includes fictional or role-play scenarios depicting such content.

1.2 Violence and graphic content

Graphic gore, depictions of torture or cruelty, instructions for creating weapons or causing physical harm, and content that encourages violence or self-harm.

1.3 Hate speech and harassment

Content that attacks, dehumanizes, threatens or incites discrimination against a person or group on the basis of race, ethnicity, nationality, religion, gender, sexual orientation, disability, age or other protected characteristic. Targeted harassment, stalking and doxxing are also prohibited.

1.4 Child-unsafe content

Any sexualization of minors (including AI-generated or fictional depictions), grooming, child exploitation material, and content that endangers the safety or wellbeing of children.

1.5 Deepfakes and unauthorized impersonation

Creating realistic depictions of a real person saying or doing things they did not do without their explicit authorization, impersonating a real individual or organization to deceive others, forged identity documents, and non-consensual intimate imagery. Synthetic media must be clearly labeled as such.

1.6 Copyright and trademark infringement

Reproducing copyrighted text, code, images or music without permission, circumventing licensing or access controls, removing attribution or watermarks, and using third-party trademarks in a way that suggests endorsement or affiliation.

2. Additional prohibited activities

• Illegal activity under applicable law, including fraud, money laundering, unauthorized access to computer systems, and distribution of malware.
• Creating or distributing disinformation intended to manipulate civic processes, or deceptive content presented as factual reporting.
• Reselling API access, credential sharing, key pooling, or circumventing rate limits and metering.
• Attempting to extract model weights, system prompts, or other non-public assets of a model provider.
• Processing sensitive personal data (health, biometric, financial account, government identifiers) without appropriate safeguards and legal basis.
• Using the service in a jurisdiction where the underlying model provider does not permit access, or to provide access to such a jurisdiction on behalf of others.

3. Content screening and enforcement

We apply a combination of automated classifiers, keyword and pattern screening, and human review to prompts and outputs. Detections are logged with a timestamp, category, request identifier and action taken. We do not use this content to train models.

Enforcement is progressive where appropriate but immediate for severe violations. Actions include: refusing a request, removing generated content, temporary rate limiting, requiring verification, suspension, and permanent account termination. We keep enforcement records for up to 2 years so that decisions can be audited and appealed.

4. Reporting abuse

Report suspected violations to support with the request ID (if available), a description of the concern and any supporting material. We aim to acknowledge reports within 2 business days and to complete an initial review within 10 business days. Reports are handled confidentially.

If you believe your content was restricted in error, reply to the notice or contact the same address with the request ID and an explanation; we will review and reverse the action where appropriate.

5. AI disclosure

All model output is generated by third-party large language models and is labeled as AI-generated in the product. You must not remove that labeling or present generated content as human-authored where disclosure is legally required. Output may be inaccurate or biased — review it before publication or reliance.

6. Provider terms

Your use must also comply with the terms of the model provider whose model you select. Where a provider's policy is stricter than this one, the stricter rule applies. We may change the available model list at any time to remain compliant with provider terms.

7. Changes

We may update this policy as laws and provider requirements change. The current version is always available at /aup with its "Last updated" date.