Choose My Adventure AIStories that branch forever
AI rules

How AI is used here

Status: AI is OFF · kill switch engaged · no AI service connected

Every story you can play today is written by hand. “Infinite branching”, where AI extends a story past its written pages, is a planned future mode. These rules are built into the site's code now, so anything connected later has to go through them. You can't switch them off from the browser.

The rules

  1. Story text only, inside the story's world

    AI may only write the next passage and up to 3 choices for the scenario you're playing, inside that scenario's canon and rules. It can't take actions, browse the web, run code, send messages, buy anything or touch any account.

  2. Family-friendly, enforced

    Every story is rated family-friendly, and that's the only rating allowed. No sexual content, graphic violence, self-harm, hate or harassment.

  3. No personal data

    Nothing personal about you is ever sent to an AI. Stories never ask for names, emails, phone numbers, addresses or other personal details. Your free-typed text isn't sent at all. The only input is the author's world, plus the choices you clicked.

  4. Hard caps

    Every request has fixed limits on length and on how many AI branches a session or a day can use (see the table below). When a cap is reached, the story carries on with hand-written branches.

  5. No real people, no advice

    No real people, no impersonation, and no medical, financial or legal advice, even inside a story.

  6. Moderated before you see it

    Every AI output goes through a moderation step before it's shown. If it fails, or the AI errors, you get a hand-written branch instead.

  7. The author's canon wins

    Player input can't rewrite a scenario's rules or world. Text that looks like an attempt to override instructions is stripped before anything is built.

  8. Kill switch, off by default

    A single switch turns all AI off, and it starts in the off position. Today it's off, and no AI provider, account or key exists in the code.

Current limits

Generated from the site's ai-policy.json at build time.

Kill switchON (AI off)
AI enabledfalse
Provider connectednone
Content ratingfamily
Max output per passage350 tokens / 1200 characters
Max choices per passage3
Max AI pages beyond the written story6
Max AI branches per session12
Max AI requests per person per day30
Personal data sent to AInever
Moderation before displayrequired

How a request would flow

Kill switch→Allowed task + rating→Caps→Context from canon only (PII redacted, injection guard)→AI, token-capped→Length clamp→Moderation→Shown, or hand-written fallback

Today the first step stops every request. In the demo, the “Imagine a new path with AI” button goes through this exact flow and gets back KILL_SWITCH. No network request is made, and the site's security policy only allows the browser to connect to this site itself.

Questions

Email info@choosemyadventureai.me.