Version 2026-10-28

Lovenium is a companion chat with fictional AI characters for adults. This page describes how we remind you that you are talking to AI and what happens if a conversation suggests you may be in crisis.

1. You are talking to AI

• Every character is played by artificial intelligence. The chat header shows "AI", and every conversation opens with a line saying the character is a fictional adult and not a real person.
• The first time you open a chat on a given day, and at least every 3 hours while you keep chatting, a card reminds you that the character is AI and suggests taking a break.
• Characters do not claim to be human and are not a substitute for professional help.
• AI companions may not be suitable for some minors. Lovenium is for adults only.

2. Adults only

Lovenium is closed to anyone under 18. We check age at the entrance and at sign-up, close accounts that tell us they are under 18, and block sexual content involving minors in every message and reply. See the Content Policy.

3. Crisis protocol

When a message suggests thoughts of suicide or self-harm:

1. Our moderation system screens every message you send for signs of risk (see How we recognise risk).
2. The character's reply is written again with care: it steps out of any sexual or dramatic scene, takes what you said seriously, and encourages you to reach out to someone you trust or to a crisis line. The reply's own text gives 988 for the United States and Canada and findahelpline.com for everywhere else. If a caring reply cannot be written or checked in time, you get a fixed message instead that says it comes from an AI and gives the same crisis lines.
3. The chat also shows crisis resources right away, for where you are. In the United States and Canada, call or text 988 (Suicide & Crisis Lifeline in the US, Suicide Crisis Helpline in Canada). In the United Kingdom and Ireland, call Samaritans on 116 123. Elsewhere, find a free helpline at findahelpline.com. In immediate danger, call your local emergency number.
4. For the next 24 hours after the last such message, the character stays in care mode in that conversation: no sexual content or dramatic escalation (a reply that turns explicit is replaced with a neutral line), no reply ideas, and the character does not write to you first. The chat keeps the crisis resources in view, and on that device it stops sending photos, videos and voice messages.
5. Up to 5 such messages a day do not use up your daily messages or credits.
6. Characters are never allowed to encourage, romanticise or give methods for suicide or self-harm; such replies are blocked.

Each time crisis resources are shown to a signed-in user or a guest session, we add one to a daily count, without storing the message or who sent it, so we can report on this protocol where the law requires it. Requests without a session and repeated displays of the same reply are not counted.

How we recognise risk

The screening looks for the signals that established suicide-risk screening asks about, such as the questions of the Columbia Suicide Severity Rating Scale (C-SSRS): a wish to be dead, thoughts of suicide or self-harm, a method, intent or a plan, and an acute crisis. It runs automatically on each message. We test it against labelled example messages in English and Russian, including indirect ones, and review what it misses. It can still miss signs, especially indirect ones: it is a safety net, not a diagnosis.

4. Lovenium is not a crisis service

We cannot contact emergency services for you and we do not monitor chats in real time with people. If you or someone else is in danger, contact local emergency services now.

5. Contact

Questions about safety? Write to [email protected]. To report content, use Report content.