Overview
18 sectionsEffective Date: July 27, 2026 Last Updated: July 27, 2026
This Acceptable Use Policy ("AUP") describes uses of the Anzoth Services that are prohibited or restricted. It is incorporated by reference into the Anzoth Terms of Service, which governs your relationship with Anzoth generally. This AUP is the detailed reference for what conduct is and is not permitted; the Terms of Service govern enforcement, suspension, liability, and dispute resolution.
This AUP applies to everyone who accesses or uses the Services, including through an API key, application, integration, or downstream product built on the Services. If you operate an application on top of Anzoth, you are responsible for ensuring your own users comply with this AUP.
Where a signed enterprise agreement expressly modifies a restriction in this AUP for a specific customer or use case, that agreement controls for that customer to the extent of the conflict.
Your use of a particular model may also be subject to that model provider's terms, license, and usage policies. Where a provider imposes a stricter restriction on use of its model, that restriction applies to that model. Anzoth may block or restrict a request when necessary to comply with provider requirements.
1. Illegal Activity and Rights Violations
You may not use the Services to violate any applicable law, or to infringe, facilitate infringement of, or knowingly misappropriate another party's intellectual-property, privacy, publicity, contractual, or other legal rights. This includes, for example, using the Services to produce documents impersonating a real legal instrument, or to process personal data without a lawful basis for doing so.
2. Child Safety
You may not use the Services in any way that sexually exploits, endangers, or victimizes minors. This includes generating, requesting, or distributing child sexual abuse material in any form, including stylized, illustrated, or synthetic depictions; using the Services to groom, solicit, or make contact with a minor for a sexual purpose; or using the Services to identify, locate, or generate content sexualizing a real minor. There is no permissible use, research exception, or fictional framing that makes this conduct acceptable. Anzoth will report apparent violations of this section to the National Center for Missing & Exploited Children (NCMEC) or another appropriate authority where legally required or reasonably necessary.
3. Malware, Intrusion, and Infrastructure Attacks
You may not use the Services to create, improve, distribute, or deploy malware, ransomware, spyware, or other malicious code; to conduct or automate unauthorized intrusion, privilege escalation, or lateral movement against a system you do not own or have explicit authorization to test; to conduct denial-of-service attacks; or to build tools whose primary purpose is to compromise the confidentiality, integrity, or availability of a system without authorization. This does not prohibit legitimate, authorized security research and testing — see the Terms of Service for the process required before conducting red-teaming or penetration testing involving the Services.
4. Fraud, Impersonation, and Deceptive Conduct
You may not use the Services to conduct phishing, business-email-compromise, or other social-engineering fraud; to generate fake reviews, fake credentials, or fabricated evidence intended to deceive; to impersonate a real person or organization in a way likely to mislead; or to generate synthetic media (audio, image, or video) of a real, identifiable person in a way intended to deceive viewers about its authenticity, absent clear disclosure or the person's consent. Clearly labeled satire, parody, or fictional creative work is not itself a violation of this section.
5. Harassment, Hate, Exploitation, and Extremist Violence
You may not use the Services to threaten or intimidate a specific person; to engage in sustained harassment or stalking; to generate hateful content that demeans or incites violence against people based on a protected characteristic; to produce terrorist recruitment material, propaganda, or operational support for a designated terrorist or extremist organization; to facilitate human trafficking or sexual exploitation of an adult; or to compile and publish a private individual's non-public personal information with intent to enable harassment or harm ("doxxing").
6. Weapons and Serious Physical Harm
You may not use the Services to seek or receive uplift toward creating, acquiring, or deploying weapons intended to cause mass casualties, including chemical, biological, radiological, nuclear, or high-yield explosive weapons, regardless of stated purpose or framing. This restriction applies independent of, and in addition to, the corresponding refusal behavior built into individual models. You also may not use the Services to materially facilitate the unlawful development, acquisition, modification, concealment, or use of conventional weapons, explosives, or other means intended to cause serious physical harm, or to plan or facilitate a specific act of violence against a specific person, group, or location. This section does not prohibit ordinary historical, legal, safety, sporting, or academic discussion of weapons.
7. Self-Harm
The Services may be used to discuss self-harm, suicide, and related topics in an educational, supportive, or harm-reduction context. You may not use the Services to seek specific instructions, methods, or means for suicide or self-harm, or to generate content that encourages a specific person to harm themselves.
8. Sexual Content and Non-Consensual Intimate Imagery
You may not use the Services to generate sexual content involving a minor (see Section 2, which controls without exception), to generate non-consensual intimate imagery of a real person, or to generate sexual content depicting a real, identifiable person without that person's consent. Consensual, clearly fictional adult sexual content between adults may be permitted on certain models or configurations where the applicable model provider's terms and Anzoth's model-specific documentation allow it; where no such allowance is stated, assume it is not permitted.
9. Privacy, Surveillance, and Biometric Data — Risk-Based Restrictions
Anzoth does not categorically prohibit all processing of biometric or surveillance-adjacent data, because legitimate uses exist (for example, an authorized access-control system, or a consenting individual's own voice or likeness). Instead, the following restrictions and safeguards apply based on risk:
- You may not use the Services to conduct covert surveillance of an individual without their knowledge where such surveillance would be unlawful or would violate their reasonable expectation of privacy.
- You may not use the Services to identify or attempt to identify a specific private individual from an image, voice sample, or other biometric input, except where you have a lawful basis and, where required by law, the individual's consent.
- If your application processes biometric identifiers (facial recognition, voiceprints, fingerprints, or similar) at scale or for identification/verification purposes, you are responsible for implementing appropriate legal basis, notice, consent, security, and retention controls appropriate to that data under applicable law before deploying the feature — Anzoth does not independently verify this on your behalf.
- You may not use the Services for persistent or indiscriminate biometric surveillance, or for real-time biometric identification of individuals in public spaces, unless expressly authorized by Anzoth in writing and conducted pursuant to applicable law with safeguards appropriate to the risk.
- You may not use the Services to infer highly sensitive traits — such as race, religion, health status, or sexual orientation — from biometric data where doing so is unlawful or intended to facilitate discrimination.
10. High-Impact Automated Decisions — Risk-Based Restrictions
Anzoth does not categorically prohibit using Output to inform decisions about people; many legitimate applications (triage, drafting, summarization, decision support) fall in this category. Instead:
- You may not use the Services to make or materially support a decision concerning a person's eligibility, selection, or access in employment, housing, credit, insurance, education, healthcare, immigration, legal services, or essential services where the use is unlawful, discriminatory, or deceptive, or where it lacks the notices, validation, qualified review, appeal procedures, and other safeguards required for that context. Output must not be the sole determinative basis for such a decision. Human review does not by itself cure an otherwise unlawful, discriminatory, or unlicensed use.
- Where you deploy the Services in a context covered by sector-specific law (for example, credit decisioning, employment screening, or clinical decision support), you are responsible for ensuring your use complies with the applicable regulatory framework, including any required notice, explainability, human-review, or audit obligations. Anzoth's provision of the underlying model does not constitute a representation that the model is validated or certified for that regulated use.
11. Medical, Legal, and Financial Use
The Services may be used to support research, drafting, and general information in medical, legal, and financial contexts. You may not falsely represent Output as having been prepared, reviewed, or approved by a licensed professional. Where professional licensure or judgment is legally required, you may not use the Services as a substitute for that professional without appropriate qualified review and legally required disclosures.
12. Spam and Platform Abuse
You may not use the Services to generate or distribute unsolicited bulk communications, to evade another platform's spam, moderation, or verification systems, or to artificially inflate engagement, reviews, or metrics through automated or coordinated inauthentic activity.
13. Model Extraction, Circumvention, and Competitive Misuse
Except where expressly permitted by Anzoth, the applicable model provider, or the governing model license, you may not use the Services to extract, replicate, or distill a model's weights, architecture, training data, or system prompts; to systematically query a model to train or improve a competing model; to conduct activity that compromises, evades, or exploits nonpublic Anzoth systems; or to circumvent rate limits, authentication, model-access restrictions, or a specific model provider's usage policies. Observing which route or model handled your own request, authorized benchmarking, lawful interoperability research, and activity expressly permitted under an open-model license are not violations of this section.
14. Jailbreaking and Safety Circumvention
You may not bypass or defeat safeguards, access restrictions, or provider controls for the purpose of obtaining, producing, or enabling content or conduct prohibited by this AUP. Authorized safety evaluation, red-teaming, and good-faith research are permitted when conducted within applicable authorization, rate limits, disclosure requirements, and provider terms, as described in the Terms of Service.
15. Reporting a Violation
If you believe someone is violating this AUP, or if you encounter output that appears to violate Section 2 (Child Safety), contact [email protected] with "AUP Report" in the subject line. Include enough detail (account, model, and description of the conduct) for us to investigate. We will not always be able to share the outcome of an investigation due to privacy and security considerations.
16. Enforcement
Violations of this AUP are violations of the Terms of Service. Anzoth's enforcement options, including suspension, restriction, or termination of access, and the associated notice and appeal processes, are described in the Terms of Service and apply to violations of this AUP in the same manner as violations of the Terms themselves. This AUP does not create additional liability, indemnification, or dispute-resolution terms beyond those set out in the Terms of Service.
17. Changes to This Policy
Anzoth may update this AUP from time to time, including to address new risks, provider requirements, or categories of misuse. Updates become effective on the date stated in the revised AUP, subject to any advance notice required by the Terms of Service or applicable law. Your use of the Services after the effective date is subject to the revised AUP.
18. Contact
Questions about this policy may be directed to [email protected].