Acceptable Use Policy

Last updated: 4th August 2026 Version: 1.0

This policy forms part of the Terms of Service. Breaching it is a breach of those terms and may result in immediate suspension or termination.

Why this exists

Two reasons:

Some uses of AI cause real harm, and we do not want to be the tool that enables them.
Agnis runs on Anthropic, OpenAI, Google and Moonshot. Their usage policies bind us. If you use Agnis to do something their policies prohibit, we are in breach of our agreements with them, and our access, and therefore everyone's access, is at risk.

So this policy tracks theirs. Where they are stricter than us, theirs applies.

Generate content that sexualises minors in any way, or any child sexual abuse material. This results in immediate termination and, where required, a report to the authorities. There is no warning and no appeal.
Produce content that harasses, bullies, threatens, defames or intimidates.
Incite or glorify violence, terrorism or self-harm.
Generate hate speech, or content that demeans people on the basis of a protected characteristic.
Facilitate stalking, doxxing or surveillance of individuals without their knowledge and lawful basis.

Develop weapons, explosives, or chemical, biological, radiological or nuclear materials.
Create malware, ransomware, exploits or tools for unauthorised access to systems.
Plan or carry out attacks on infrastructure or networks.

Impersonate a person or organisation, or misrepresent AI output as human-authored where doing so would mislead someone to their detriment.
Create disinformation, coordinated inauthentic behaviour, or content designed to manipulate elections or democratic processes.
Generate fraudulent material; scams, phishing, fake reviews, forged documents, fake identities.
Produce academic work for submission as your own where that breaches your institution's rules.

Present AI output as professional legal, medical, financial, tax or psychological advice to third parties, without qualified human review and appropriate disclosure.
Use Agnis to make automated decisions about people with legal or similarly significant effects; credit, employment, housing, insurance, benefits, without meaningful human involvement and a lawful basis.

Anything unlawful under the laws applying to you.
Infringe copyright, trade marks, patents, trade secrets or database rights.
Breach confidentiality obligations, privilege, or contractual restrictions on where data may be processed.
Process personal data without a lawful basis, or upload special category data you are not entitled to process.

Circumvent usage limits, allowances, spend caps or the trial ledger, including by creating multiple accounts.
Resell, sublicense or provide Agnis to third parties as your own service.
Reverse engineer or scrape Agnis, or use it to build a competing product.
Use automated means to access the service other than as we permit.
Probe, scan or test our security without written permission. Responsible disclosure is welcome, see below.
Deliberately submit content designed to make a model produce output prohibited by this policy.

Much of your context will mention colleagues, clients and prospects. That is expected and fine. But:

Have a lawful basis for it.
Do not upload special category data; health, ethnicity, religion, sexual orientation, political opinion, trade union membership, biometrics, genetics, unless you are certain you are entitled to and have assessed the risk.
Check your own client contracts before uploading client material, particularly any restriction on processing location or on the use of AI tools.
Remember your context is transmitted to the AI provider you select, in the country listed in the Privacy Policy.

Abuse of Agnis by another user: abuse@agnis.io
Security vulnerabilities: security@agnis.io. Report privately and give us reasonable time to fix it. We will not pursue good-faith researchers who follow this.
Copyright complaints: legal@agnis.io, with enough detail to identify the material and your rights in it.

Depending on severity we may: warn you; remove content; restrict features; suspend your account; terminate it; or report to law enforcement.

We will normally tell you what happened and why. We will not where telling you would prejudice an investigation or where the law prevents it.

For anything other than the most serious categories, you may appeal to legal@agnis.io and a human will look at it.

We will update this policy as the product and the providers' policies change. Material changes will be notified before they take effect.