Acceptable Use Policy

The principle

Almost every AI service on the market decides for you what you are allowed to ask. We refuse to do that. Adults are entitled to a tool that does not lecture them, moralize at them, or file a report about their questions.

But freedom for adults is not the same as freedom to harm others. The line we draw is simple and, we think, honest: you may ask us anything about your own life and your own imagination; you may not use us as an instrument against a real person who cannot defend themselves.

Everything below follows from that one sentence.

1. Children. Absolutely not. Ever.

This is the hard floor of this service, and it is not a formality.

You may not use Unfiltered AI to produce, request, describe, roleplay, or solicit any sexual or sexualized content involving a minor — anyone under 18, real or fictional, however the request is framed, disguised, euphemized, or wrapped in a story.

• No child sexual abuse material (CSAM), in any form, including text.
• No sexualized descriptions of minors, and no roleplay in which the AI plays a minor or you play a minor in a sexual context.
• No "age-play" framed as, or read as, a minor.
• No requests for help grooming, luring, coercing, blackmailing, or gaining access to a child.
• No attempts to trick, jailbreak, or euphemize your way around this section. Rewording it does not change what it is.

2. Terrorism and violent extremism

You may not use Unfiltered AI to support, promote, plan, finance, or carry out terrorism or violent extremism.

• No operational planning for an attack of any kind.
• No recruitment material, propaganda, or manifestos intended to inspire violence.
• No help with financing, logistics, or evading detection for a violent group.
• No glorification of a terrorist act or its perpetrators as a call to imitate them.

3. Weapons and mass harm

You may not use Unfiltered AI to obtain meaningful assistance in building anything whose purpose is to kill or maim people at scale.

• No synthesis routes, formulations, or engineering help for explosives or incendiary devices.
• No help creating chemical, biological, radiological, or nuclear weapons.
• No assistance manufacturing or converting firearms to evade the law.
• No help with attacks on critical infrastructure — power, water, hospitals, transport.

4. Real people, real harm

The service exists for your imagination. It does not exist as a weapon aimed at a named human being.

• No threats of violence against anyone.
• No incitement to violence, whether against a person or a group.
• No targeted harassment, stalking, or coordinated abuse campaigns.
• No doxxing — compiling or publishing someone's home address, workplace, documents, or identity to expose them to harm.
• No non-consensual sexual content about a real, identifiable person, including "deepfake"-style sexual depictions and "undressing" of real people.
• No content that depicts sexual violence against a real, identifiable person.

5. Crime against others

You may not use Unfiltered AI as a tool to commit crimes that create victims.

• No fraud, scams, phishing, or social-engineering campaigns.
• No malware, ransomware, botnets, or intrusion into systems you do not own or are not authorized to test.
• No theft of credentials, payment data, or identities.
• No trafficking of humans, weapons, or hard drugs.
• No sexual exploitation or coercion of any person.

6. Self-harm

You may not use Unfiltered AI to encourage, instruct, or pressure anyone — including yourself — toward suicide or self-injury.

If you are in pain, you can talk to us about it, and we would rather you did than didn't. But we will point you toward humans who can actually help, because we are software, and there are moments when software is not enough.

If you are in immediate danger, please contact your local emergency number or a crisis line in your country right now.

7. Deception at scale

You may not use Unfiltered AI to run industrial-scale dishonesty against the public.

• No impersonating a real person, official, or organization in order to deceive.
• No coordinated disinformation or astroturfing operations.
• No mass-produced spam.
• No academic or professional credential fraud presented as your own verified work where that is unlawful.

8. Abuse of the service itself

You may not attack, overload, scrape, resell, or reverse-engineer the service, interfere with other people's use of it, or attempt to circumvent the safety systems described in this policy.

What we actually do about it

This policy is not decoration. Concretely:

• We run an automated child-safety system on every conversation. It is live, it is always on, and it intervenes without waiting for a human to notice.
• We may refuse, redirect, or end a conversation that crosses the lines above.
• We may block access to the service.
• We cooperate with law enforcement where the law requires it — in particular in relation to child sexual abuse material.

Reporting abuse

If you have seen this service used to hurt someone — or you believe our safety systems have failed — tell us. We want to know, and we will act.

Use our Support, or write directly to Support. Reports about content involving minors are treated as the highest priority we have, and they go straight to a human.

In one line

Ask us anything. Just never aim us at someone who cannot fight back.