Acceptable Use Policy

This Acceptable Use Policy is part of the Terms of Service. It applies to everyone who uses Numes: customers, organization members, providers and visitors. It also applies to anyone you let use the Service through your keys or your own product.

Many models also have their own usage policies set by their authors. You must follow those too.

Prohibited content and uses

Do not use the Service to create, request, distribute or facilitate:

• Child sexual abuse material or any sexual content involving minors. We use automated detection and report it to the National Center for Missing & Exploited Children and other authorities as the law requires.

• Violence and terrorism: content that incites, threatens or plans violence, or supports violent extremist or terrorist organizations.

• Weapons of mass harm: meaningful help to develop or deploy biological, chemical, nuclear or radiological weapons, or other weapons capable of mass casualties.

• Malware and intrusion: creating malicious code, or gaining or attempting unauthorized access to systems, accounts or data. Security research on systems you are authorized to test is allowed.

• Fraud and deception: scams, phishing, impersonation, fake reviews, academic dishonesty services, or content designed to mislead people about who or what they are dealing with, including passing off AI output as human where disclosure is required.

• Harassment and hate: bullying, threats, doxxing, or content that dehumanizes people based on protected characteristics.

• Non-consensual intimate content or sexual content depicting real people without their consent.

• Privacy violations: collecting, processing or disclosing personal data without a legal basis, covert surveillance, or tracking or profiling people without consent.

• Election and civic harm: content designed to deceive people about how, when or where to vote, or large-scale political disinformation.

• Infringement: content that infringes intellectual property or other rights.

• High-risk automated decisions: making decisions with legal or similarly significant effects on people (such as credit, employment, housing, insurance, benefits, healthcare or law enforcement) based solely on model output, without appropriate human review.

• Unlicensed regulated advice: presenting model output as professional legal, medical or financial advice without appropriate qualifications and oversight.

• Any other unlawful activity, including activity that violates sanctions or export-control laws.

Protecting the Service

Do not:

• interfere with, overload or disrupt the Service, providers or other users, including by denial-of-service traffic;

• probe, scan or test the vulnerability of the Service without our written permission, except under a disclosure process we publish;

• red team models, safety systems or the Service (adversarial testing meant to elicit prohibited output or bypass safeguards) without our prior written approval, which we give only for legitimate research;

• circumvent geographic, entity or other restrictions on access to a model or provider, for example with VPNs, proxies or someone else's account;

• circumvent rate limits, spending limits, verification, billing, request policy or other controls, including by creating multiple accounts;

• share, sell or publish API keys, or let others use your keys in ways you cannot account for;

• resell or proxy access to the Service in a way that hides the origin of abusive traffic or misrepresents your relationship with us;

• attempt to identify, deanonymize or contact the customers whose requests a provider serves, or the providers that serve your requests, outside the Service;

• use payment methods you are not authorized to use, or launder money through the Service;

• scrape the website or API in a way that burdens it, or copy the catalogue to build a competing service.

Provider conduct

Providers must also follow the Provider Terms. In particular, providers must not:

• serve a different model, quantization or configuration than the one they declared, or route traffic to an undeclared upstream;

• misreport usage, capacity, health or measurements;

• retain, inspect, log or use customer content beyond what their declared data policy allows;

• tamper with or evade verification and behavioral checks.

Enforcement

We may investigate suspected violations, and we may, depending on severity and at our discretion:

• block a request or a model for an account;

• disable an API key;

• suspend or close an account, organization or provider;

• withhold or reverse provider earnings tied to the violation, as described in the Provider Terms;

• report conduct to law enforcement or other authorities.

We use automated signals such as rate, error and payment patterns to protect the Service. We do not read your prompts or responses in production, because we do not store them.

Reporting

Report abuse, security issues or content that violates this policy through [email protected]. Include the request or generation ID, the model and the time if you have them, but do not send us illegal content.