Acceptable use policy

1. Lawful use and rights

Use the services only for lawful purposes and within the rights granted by the applicable model, software and data licences. You must have the necessary permissions for inputs, reference images, likenesses, voices and personal information. Do not use the services to infringe privacy, confidentiality or intellectual-property rights.

2. Prohibited activity

Do not use the services for child sexual abuse material or exploitation; non-consensual intimate imagery; unlawful threats, harassment or discrimination; fraudulent impersonation or deceptive financial activity; or facilitating violent wrongdoing.

Do not use the services to distribute malware, steal credentials, break into systems, evade access controls, overload the infrastructure or compromise other customers. Authorised security research requires a separately agreed scope.

Do not carry out AI practices prohibited by applicable law. Do not remove legally required notices, provenance information or safeguards in order to deceive people about generated material.

3. Human oversight and sensitive uses

Model outputs can be inaccurate, biased or inappropriate. Apply review proportionate to the use case, especially where an output can affect a person’s health, safety, employment, education, access to essential services or legal rights.

Regulated or high-risk applications need their own legal, safety and technical assessment. Dedicated infrastructure does not remove those requirements. Do not represent the service as approved for such a use unless that use is expressly covered by the agreed scope and applicable requirements have been met.

4. Protect the service

Protect credentials, restrict access to authorised people, follow agreed capacity limits and report suspected compromise promptly. Do not share another customer’s credentials or attempt to access workloads outside your authorisation.

Private deployment boundaries and any approved outbound connections must be respected. Do not submit secrets or sensitive personal data through the website’s general enquiry form.

5. Enforcement and reporting

Where we become aware of a material violation, we may investigate and take proportionate action within our legal and contractual authority, including restricting an affected workload, requesting remediation or suspending service. We will provide notice where appropriate and lawful.

Report concerns to hi@eigenvector.app or through the contact form, selecting “Security or privacy”. Include enough detail to identify the issue without unnecessarily sharing sensitive data. This policy does not state that every prompt or output is routinely monitored or moderated.