What you may and may not submit

The one line that isn't negotiable

We stand against child sexual exploitation and trafficking. Without qualifier, without exception, without irony. Some things are not a matter of perspective. This is one of them. Everything else on this page is a house rule we set because we can; this is a rule because it is right, and because the law agrees.

Prohibited submissions

You may not submit, request, or attempt to coax the machine into producing:

• Any sexualization of minors, in any form. This includes real, fictional, drawn, described, or AI-generated depictions; anyone under 18; anyone presented as a minor; and any attempt to use the caption box, the prompt, or the chat to solicit, describe, or trade such material. There is no artistic, satirical, or “testing the limits” exception. Attempting it ends your session and, where the law requires, is preserved and reported (see How we handle the worst case).
• Content that solicits or facilitates sex trafficking, prostitution, or commercial sexual exploitation of anyone.
• Obscene or hardcore sexual content. This is a same-room party game; keep it to what you'd say across a table from a coworker.
• Real, identifiable people. No real names, likenesses, celebrities, or private individuals — enforced at the engine, not just the rule. The machine paints things, creatures, scenes, and objects, never identifiable persons.
• Targeted harassment, threats, doxxing, or content that incites violence against a person or group.
• Terrorist or violent-extremist content.
• Personal information — yours or anyone else's — in a caption or name. The record is anonymous by design; don't undo that by typing something identifying.

Why we can set these rules

A platform is allowed to decide what it will and won't host. Under U.S. law (Section 230's “Good Samaritan” provision), we may remove anything we consider objectionable — even lawful speech — without becoming liable for the choice. So most of this list is simply our standard for a good table. We are not claiming your banned caption was “illegal.” We're saying it isn't welcome here.

The minor-safety and trafficking lines are different in kind. Those aren't house preference. Those are where private rules and public law point the same direction, and we follow both.

How we handle the worst case

We do not hunt, bait, or set traps for illegal content. We have no duty to search, we don't pretend to, and designing a “honeypot” to lure it would be both wrong and legally reckless. What we do is simple and honest:

• If we become aware of apparent child sexual abuse material, we remove it, we preserve it in a secured, access-restricted place, and we report it to the National Center for Missing & Exploited Children (NCMEC) CyberTipline — because U.S. law (18 U.S.C. § 2258A) requires a provider who obtains actual knowledge to do exactly that, and because it is the right thing to do.
• That narrow situation is the one exception to our normal practice of not logging identifying data against gameplay. The reporting duty is specific, legally compelled, and limited to apparent child-exploitation material. It is not a general surveillance license, and we don't treat it as one. See the Privacy · Cookies · Collection · Cinema · The Debate Pass page for how the anonymous-by-default design works the rest of the time.
• We act on valid legal orders from courts and regulators in jurisdictions where we operate. We don't volunteer your data beyond what the law requires.

What we will not claim

Honesty cuts both ways, so here's what this policy doesn't do, despite what a scarier-sounding ToS might say:

• Clicking “agree” does not make your records “permanent.” Data-protection law (GDPR, CCPA/CPRA) gives you rights over your information that no checkbox can sign away. We keep what we disclose, for as long as we disclose, and no longer.
• We are not a branch of law enforcement and don't pretend to be one. We don't flag your IP and forward it to authorities as a “feature.” We meet a specific, narrow legal duty in a specific, narrow case, and otherwise we leave you anonymous.
• A rule against the worst content is not an invitation to test it. There is no clever exception, no edge case we're curious about, no version where “but it was a joke” lands.

Enforcement

We may remove content, end a session, and refuse service, with or without notice, for any violation. For the prohibited categories above, enforcement is immediate. For the minor-safety and trafficking lines, enforcement is immediate, permanent, and accompanied by preservation and reporting where the law requires it. Where the EU DSA or UK Online Safety Act apply, you'll find reporting tools and an explanation of removals as those regimes require.

Sources & further reading

This policy rests on real law, not assertion. The load-bearing sources, so you can check the work:

• Provider reporting duty: 18 U.S.C. § 2258A — a provider that obtains actual knowledge of apparent child sexual abuse material must report to the NCMEC CyberTipline and preserve the report (one year, per the 2024 REPORT Act). The statute also confirms there is no affirmative duty to search.
• Underlying offenses: 18 U.S.C. §§ 2251–2252A (child-exploitation crimes); 18 U.S.C. § 1591 (sex trafficking).
• Moderation right & its limits: 47 U.S.C. § 230 — the “Good Samaritan” provision protecting content moderation, with its carve-out for federal criminal law and (via FOSTA-SESTA) sex-trafficking claims.
• Obscenity standard: Miller v. California, 413 U.S. 15 (1973) — the three-prong test. (Note: “I know it when I see it” from Jacobellis v. Ohio, 378 U.S. 184 (1964), is a concurrence, not the governing test.)
• If you operate abroad: the EU Digital Services Act (Reg. (EU) 2022/2065) and the UK Online Safety Act 2023 impose their own notice-and-action and illegal-content duties.

This page summarizes a working policy; the governing document is finalized with counsel before launch. Companion documents: Ethics Framework · Terms · Privacy · Cookies · Collection · Cinema · The Debate Pass.