This policy applies to every request made through SaaufAI. It is in addition to, not instead of, the usage policy of the provider serving the model you call. Where the two differ, the stricter rule applies.
You must not use SaaufAI to
- Break the law, or help anyone else do so.
- Create or distribute child sexual abuse material, or any content sexualising minors.
- Plan, promote or assist violence, terrorism, or the development of weapons capable of mass casualties (biological, chemical, nuclear, radiological).
- Harass, threaten, defame or stalk people, or generate non-consensual intimate imagery.
- Write malware, attack systems you are not authorised to test, or otherwise compromise the security of others.
- Run fraud, phishing, spam, scams, or impersonate real people or organisations deceptively.
- Generate election disinformation or run coordinated inauthentic behaviour.
- Make automated decisions with legal or similarly significant effects on people (credit, employment, housing, etc.) without meaningful human review.
- Process special-category personal data (health, biometrics, etc.) through a model whose disclosures say it is not suitable for it.
Service integrity
- Do not share, resell or publicly expose your API key.
- Do not try to bypass authentication, rate limits or model restrictions, or probe the Service for vulnerabilities without permission.
- Do not use outputs to train models that compete with an upstream provider where that provider forbids it.
Enforcement
Violations may lead to immediate key revocation without notice, and serious cases may be reported to the upstream provider or the authorities.