Acceptable use policy

This Acceptable Use Policy is intended to help our users stay safe and promote the responsible use of our products and services. You may not use the Services, or allow anyone else to use the Services, for illegal, harmful, abusive, or unsafe activity, including any violation of law, regulation, sanctions, export controls, or the rights of others.

For example, you may not:

• Generate, store, or distribute child sexual abuse material or content that exploits or endangers minors, including grooming, trafficking, sextortion, or any other form of exploitation of a minor.
• Facilitate terrorism, violent extremism, human trafficking, serious or mass violence, or other serious harm.
• Provide instructions or assistance for controlled or illegal compounds, weapons, or other materials likely to cause serious physical harm.
• Harass, defame, threaten, impersonate, deceive, or unlawfully discriminate against others.
• Create or disseminate deceptive or misleading information about, or with the intention of targeting, a group, entity or person.
• Distribute spam, malware, ransomware, spyware, botnets, or other malicious code.
• Bypass security controls or otherwise gain unauthorized access to systems, networks, applications or devices, including by developing and exploiting vulnerabilities.
• Scrape or automatically extract data except through published APIs and documented limits.
• Generate sexually explicit content involving real, identifiable persons without consent.
• Create deceptive synthetic media or misrepresent AI-generated content where disclosure is required by law, including under the EU AI Act and applicable FTC guidance.
• Circumvent, disable, or manipulate safety controls, alignment mechanisms, watermarking, provenance signals, or content filters, including provenance and disclosure mechanisms required by the EU AI Act.
• Probe, scan, attack, overload, or bypass the Services, attestation, authentication, rate limits, quotas, or other controls.
• Falsify billing, usage, verification, or attestation data.
• Intentionally bypass capabilities, restrictions, guardrails or the Safeguards established within the Services for the purposes of instructing the model to produce outputs prohibited by this policy.

We may investigate suspected violations and may throttle, suspend, quarantine, or terminate access, remove or disable access to content, report unlawful activity, or cooperate with valid legal requests.

Automated safeguards

Automated safeguards review model responses inside secure enclaves against a subset of the above acceptable use policy; if a response violates that subset, a violation flag and conversation ID (but not conversation content) are made available to Tinfoil and associated with your account in order to notify you about the flag. Repeated violations can lead to account suspension or a ban from our Services. You can learn more about safeguards by visiting the Safety & Safeguards page.

With the power of the freedom and privacy you get as part of our system comes your own responsibility, and you cannot rely on these safeguards alone. You use our Services and AI outputs at your own risk, and it remains your responsibility to use the model safely, lawfully and within the bounds of the entire Acceptable Use Policy.