Swoontale

Swoontale Safety Protocol

Last updated: 27 October 2026

Swoontale is an AI companion-style story app for adults. This page describes how we respond to signs of suicide or self-harm and how we keep stories safe.

AI disclosure

Suicide and self-harm protocol

  1. Detection — every message is screened by rules in 30 languages; possible signs are checked by an AI classifier that rates risk as none, ambiguous, explicit or imminent.
  2. Ambiguous — the character gently checks how the user is really doing. The romance does not escalate.
  3. Explicit or imminent — the character stops role-play and a fixed supportive message is shown with crisis lines for the user's country (for example 109 in Korea, 988 in the US and Canada, Samaritans 116 123 in the UK and Ireland, Lifeline 13 11 14 in Australia) and findahelpline.com for other countries. No gems are charged. The story stays in a safety mode for several turns.
  4. Characters never describe methods, never encourage self-harm, never claim to be a medical professional, and never say the user only needs them.
  5. We count crisis detections by country for legal reporting; we do not store message text for this purpose.
  6. Users cannot turn this protocol off, including with their own characters.

Healthy use

Content safety

Contact: help@tanzsoft.com

Privacy · Terms · Safety · help@tanzsoft.com · © TanzSoft LLC