Acceptable Use Policy使用规范

This policy applies to everyone who uses the API, directly or through a product built on it. Wholesale customers must enforce it with their own users. It supplements the Terms of Service.

1. Content you may not create or distribute

• Sexual content involving minors, in any form. Accounts are closed immediately and reported where the law requires.
• Sexual imagery of real people without their consent, and non-consensual intimate imagery of any kind.
• Content that impersonates a real person or organisation in order to deceive: synthetic video or audio of real people presented as genuine, fake endorsements, fraudulent political or electoral material.
• Incitement or glorification of violence or terrorism; instructions for weapons capable of mass harm.
• Harassment, threats, or hateful content targeting people for who they are.
• Malware, phishing, scams, spam, or any content used to defraud.
• Content that infringes someone else's copyright, trademark, privacy or publicity rights.
• Content that is illegal where you or your users are, or that breaches the usage policies of the upstream model provider. Requests are processed in China; the upstream may refuse some requests under its own rules, and such refusals are passed through to you unchanged and are not billed.

2. Things you may not do

• Share, publish or resell your keys, or run a public or "free" proxy on top of them. Reselling access is allowed only under a written wholesale agreement.
• Circumvent rate limits, billing, safety filters or account restrictions, including by creating multiple accounts for free credit.
• Probe, scan or attack the service or the upstream provider, or send traffic designed to overload them.
• Use outputs to train or improve a competing model where the upstream provider's terms prohibit it.

3. Labels and disclosure

Keep any watermark or provenance metadata that the model adds. Disclose that content is AI-generated where the law of your market requires it (for example China's labelling rules for synthetic media and the EU AI Act's transparency duties), and do not present generated media as a recording of real events.

4. Enforcement

Depending on severity we may rate-limit a key, suspend it, or close the account. Serious cases (section 1, first three items; attacks) lead to immediate closure. Where an account is closed for a breach, we may withhold the remaining balance to the extent the law allows. We cooperate with lawful requests from authorities. Wholesale customers must suspend end users who breach this policy when we ask, within 24 hours.

5. Reporting

To report abuse, use the contact details shown on this site.