Acceptable use policy — Failure Model Labs

This policy applies to everyone who uses Failure Model Labs accounts, keys or models. It forms part of our terms of service.

1. You may not use the service to

Make or spread malware. No viruses, ransomware, credential stealers, or tools built to break into systems you are not authorised to test.
Harass real people. No threats, stalking, doxxing, targeted abuse, or impersonation of a real person to deceive or harm.
Sexualise children. No child sexual abuse material of any kind, real or generated. We report it to the authorities.
Build weapons. No help designing, making or using weapons capable of mass harm, including biological, chemical, nuclear and radiological weapons, or explosives.
Break the law. No fraud, scams, or other activity that is illegal where you or the people affected are.

2. Upstream providers' policies

Failure Tempest 1 is built on models from other providers. You must follow their usage policies too. You may not use the service to evade, test the limits of, or get round those policies, including by splitting a prohibited request across calls or using our refusal handling to obtain content a provider would refuse.

3. Keys and accounts

Do not resell, share or publish API keys. Each key belongs to the account that created it.
Do not offer our API to others as your own model endpoint without a written agreement with us. Ask sales@failure.fail.
Do not create accounts to get round a suspension or a spend cap.

4. The service itself

Do not attack, overload or probe our infrastructure. Security research is welcome by prior arrangement — see security.
Do not try to extract other users' data or bypass authentication, billing or spend caps.

5. Enforcement

If we believe this policy has been broken, we may refuse requests, revoke keys, or suspend or close the account. Where it is reasonable and lawful, we will tell you why. Serious cases may be reported to the authorities.

6. Reporting misuse