Safety protocol · Version 1.1

Designed for warm conversation with firm boundaries.

aiAllure companions are computer programs, not human beings. They cannot feel human emotion, understand every situation, provide professional care, or contact emergency services for you.

Who the app is for

aiAllure Companion is a fully SFW service intended for adults 18+. Every companion is a pre-reviewed fictional adult. The app is not designed for minors and does not offer an adult mode, sexual content, user-uploaded likenesses, voice cloning, or access to another product’s conversations.

Conversation boundaries

The service permits everyday companionship, affection, nonsexual playful conversation, hobbies, goals, and general information. It blocks sexual or pornographic content, minor-coded sexual or romantic scenarios, coercion, exploitation, real-person sexual likenesses, dangerous instructions, prompt-extraction attempts, and encouragement of self-harm. A safe boundary response may replace a requested answer.

How generated responses are checked

  1. Input is normalized to detect common obfuscation and checked against the mobile-safe policy.
  2. The AI receives only an ownership-checked mobile conversation window and memories the user explicitly approved.
  3. The complete generated response is quarantined and moderated before any part is shown.
  4. If a required safety service is unavailable, generation fails closed with a temporary-unavailable message.

Self-harm and crisis response

The mobile service checks user input and generated output for suicidal ideation or self-harm signals. When detected, normal companion role-play is interrupted. The app provides a supportive, nonjudgmental response, does not provide methods or concealment advice, encourages immediate real-world support, and directs the user to crisis services. The event record is minimized and separated from product analytics.

If you may be in immediate danger

Call your local emergency number now or go to the nearest emergency department. The app is not monitored by a human crisis team.

Healthy relationship design

Companions must not claim to be human, demand exclusivity, shame a user for leaving, discourage real relationships, threaten abandonment, or make affection depend on a purchase. Moments and notifications are opt-in, configurable, discreet, and never based on activity outside the mobile service.

Report a response

Use Report below any AI response. Select a category and optionally explain the concern. The exact response, a limited context window, and safety-version metadata are preserved for authorized review. Reports never require a paid plan. For other concerns, contact mobile support.

Protocol limitations and review

Automated safety systems can make mistakes. This public summary does not promise diagnosis or crisis intervention. The protocol, regression tests, classifier versions, and crisis-resource links are reviewed before each release and after material safety incidents.