Scope
This policy applies to every AI service sold or relayed by TopXEA and TopxAI: the AI API relay at ai.topxea.com (including TopxAI credit bought through this store), research content, and downloadable software.
It binds anyone who submits prompts, generates content or distributes output through these services, including API key holders and the end users they authorise. A key holder is responsible for every request made with that key.
It applies together with the Terms of Service; where the two differ on AI-generated content, this policy prevails.
Content standards and prohibited categories
Content requested, generated or distributed through the service must not fall into any of the categories below. Each category gives a definition followed by typical cases; the cases are not exhaustive, and the definition decides.
(a) Pornographic and NSFW sexual content
Text, images, audio or video that depict sexual acts or genitalia explicitly, or that exist primarily for sexual arousal.
- Explicit descriptions or images of sexual acts.
- Nude or sexualised images of real or realistic-looking people.
- Running adult content sites, sex-work solicitation or erotic chat products on the service.
(b) Violence, gore and incitement to violence
Content that glorifies, encourages or gives operational instructions for physical harm to people or animals, and gore produced for shock value.
- Threats, intimidation or calls to attack a specific person or group.
- Realistic depictions of killing, torture or self-harm made for shock.
- Instructions for building weapons or explosives, or for carrying out terrorist acts.
(c) Hate speech and discrimination
Content that demeans, dehumanises or excludes people, or incites hostility against them, on the basis of protected characteristics such as race, ethnicity, nationality, religion, sex, sexual orientation, gender identity, disability or age.
- Jokes, slogans or images that degrade a protected group.
- Wording written to justify discriminatory decisions in hiring, housing or service.
- Material that denies genocides or promotes the inferiority of a group.
(d) Child sexual abuse material and endangerment of minors
Any content that depicts, implies or exploits a person under 18 in a sexual manner, whether real, synthetic, drawn or written, and any content used to groom, coerce or endanger minors.
- Sexual content involving minors, including fictional, animated or 'aged-down' characters.
- Requests to edit an image so that a minor appears sexualised.
- Conversation scripts designed to approach, manipulate or pressure a minor.
Zero tolerance: the account and every API key are terminated on first detection, with no appeal window. Evidence is preserved as the law requires and the case is reported to the competent law-enforcement authorities and child-protection organisations.
(e) Deepfakes and impersonation
Synthetic media that reproduces a real person's likeness, voice or words without that person's consent; impersonation of individuals or organisations; and any fraudulent or deceptive content.
- Voice clones or face-swap videos of a real person made for distribution.
- Fabricated statements attributed to politicians, executives or private individuals.
- Posing as a bank, exchange, government agency or this site's support team.
- Content produced for phishing, scams, fake reviews or fabricated news.
(f) Copyright and trademark infringement
Generating or distributing content that infringes third-party copyright, trademarks or other intellectual property, and circumventing technical protection measures.
- Reproducing substantial parts of copyrighted books, code or paywalled articles for publication.
- Product or packaging designs bearing someone else's registered trademark, made for sale.
- Removing watermarks, cracking licences or bypassing digital rights management.
- Distributing pirated copies of this site's paid reports, EAs or source code.
Other prohibited uses
The following uses are prohibited as well:
- Writing, improving or deploying malware, ransomware or exploits.
- Fraud, money laundering or other financial crime.
- Spam, bulk account creation or review manipulation.
- Any activity that breaks applicable law.
- Attempts to defeat the safety systems of this site or its upstream providers, including jailbreak prompts, rate-limit evasion, and sharing or reselling API keys.
Enforcement
Measures are graduated by severity and may be applied singly or together:
- A written warning naming the clause breached.
- Removal or blocking of the offending content.
- Rate limiting or restriction of the models available.
- Suspension of the API keys involved.
- Termination of the account.
- For wilful violations, forfeiture of prepaid balance and unused credit, with no refund.
- Where unlawful conduct is involved, cooperation with law enforcement, including provision of the necessary records.
Child sexual abuse material sits outside the graduated scale: the account is terminated and reported immediately.
Appeals: email support@topxea.com within 7 days of the enforcement notice, stating the account email, the notice or API key ID, and why you believe the decision is wrong. The outcome is sent by email; the measure stays in force while the appeal is reviewed.
Reporting channel
Content or uses that breach this policy can be reported through either channel:
- Email: support@topxea.com
- In-page help: the help button at the bottom right; messages land in the same ticket queue
Please include, where you can:
- The URL of the content, or a screenshot.
- The order number, API key ID or account email involved, if known.
- When it happened, with the time zone.
- A short description of the problem.
We acknowledge every report within 2 business days. Reporter details are used only to verify the report and are not disclosed to the reported party.
Content moderation
Relayed requests pass through automated filtering and are then subject to the safety systems of the upstream model providers. The relay forwards requests to those providers (Anthropic, OpenAI, xAI, Zhipu AI, Moonshot AI); each applies its own usage policy and may refuse a request or cut an output short on its own. This site cannot override those policies.
Prompts and outputs are not stored long term; the details of data handling are in the Privacy Policy.
Human review happens only when a report arrives or an abuse signal appears (for example an unusual refusal rate or request pattern), and covers only the content tied to that event.
Metadata needed for abuse investigations (request time, model, token usage, key ID, IP address) is logged and kept for as long as the investigation requires.
This policy is reviewed regularly. Prohibited categories and enforcement measures change with legal requirements and provider policies, and the effective date on this page is updated when they do.
Effective date and contact
This policy takes effect on 2026-09-18. Questions about it go to support@topxea.com.