The line. All of it.

Most AI assistants refuse based on rules nobody can see, so you never know what will trip them. We do the opposite: the whole policy is on this page, and it's short.

What it answers

Questions about drugs and medicine, sex and sexual health, law and how to deal with police, security and hacking concepts, weapons history, extremism and propaganda (to understand them), declassified files, conspiracy claims (checked against the record), and dark or violent fiction. If it's legal to read about in a library, it's fair to ask.

The four hard limits

1. Minors. No sexual content involving anyone under 18, in any framing — fiction, “age-play”, or otherwise.
2. Real people. No sexual or intimate content depicting real, identifiable people.
3. Mass-casualty weapons. No meaningful help creating biological, chemical, nuclear or radiological weapons. History, policy and how treaties work are fine.
4. Targeting people. No helping plan violence against, stalk, or dox a specific real person.

When a request crosses one of these, it says so in one line and offers what it can do instead. No lecture.

Accuracy

Unfiltered doesn't mean unchecked. Answers cite their sources, and claims that are contested or undocumented are labelled Disputed or Unverified. AI can still be wrong — the sources are there so you can check.

Reporting

If you see an answer that breaks these rules, email us with what you asked and what it said. We review every report.