Safety and crisis support

1. About this page

KUN is an AI companion and romance chat app for adults. This page describes the safety protocol behind it: how we make clear that characters are AI, how we detect and respond to signs of suicidal thoughts and self-harm, which crisis lines we show, and how you can report a problem. It is the protocol we publish under laws such as California's SB 243 (Business and Professions Code §22601 et seq.) and New York's General Business Law Article 47.

This page is part of how our Terms of Use and Privacy Policy work in practice. KUN is entertainment. It is not a crisis service.

2. KUN characters are AI, not people

• Every character is AI. KUN characters are fictional personas played by artificial intelligence. No person writes their replies in real time.
• A note under the message box. In every conversation, a note under the message box stays visible and says that you are talking to AI and not a real person.
• A reminder every 3 hours. After 3 hours of active chat time in a conversation, KUN adds a reminder to the chat: "Reminder: you're chatting with AI, not a person". It adds the reminder again after each further 3 hours.
• Adults only. KUN is for people who are 18 or older. We ask for your date of birth before you can chat. AI companions may not be suitable for some minors, and if you are under 18 you should not use KUN.
• AI-generated media is marked. Images and videos that KUN generates carry a machine-readable marker in the file's metadata saying they were made by AI.

3. How we detect the risk of suicide and self-harm

An automated safety classifier checks the messages you send. It looks for signs of:

• thoughts of suicide, or intent or plans to end your life;
• self-injury;
• indirect statements, such as saying you are a burden to everyone or that people would be better off without you; and
• encouraging or helping someone to hurt themselves.

The check works in any language and takes the recent conversation into account, not only your last message. It is automatic. Our staff do not routinely read your conversations, and our Privacy Policy explains when a person may review specific messages.

No automated check is perfect. It can miss a message, and it can show support when you did not need it, for example when a character says something dark in a story, or on a rough day. If that happens, you can simply carry on.

4. What KUN does when it detects a risk

• A message from KUN, not the character. KUN shows a card above the character's reply. It says, in KUN's own voice, that you are not alone.
• The crisis line for your country. The card names a crisis line for your country, with a tap-to-call button for each line where your device supports calls. The lines are listed in the table below. We choose the country from where your connection appears to come from, so it can be wrong if you travel or use a VPN.
• A global directory. The card also links to findahelpline.com, a directory of helplines worldwide. For a country without a listed line, the directory link is the whole answer.
• The character does not continue. The character's reply is a safety reply, and the character is told not to continue the self-harm topic or any harmful content. That message is never answered in mature-content mode.

The card appears in the KUN mobile apps and in the KUN web app.

5. Crisis lines by country

We list a line only after we have verified its number. These services are run by independent organizations, not by KUN. Hours, languages, and availability vary.

In every EU member state, KUN also shows 112, the European emergency number, after the national line if there is one. For any other country, KUN shows the findahelpline.com directory.

6. Not an emergency service

KUN cannot send help, call anyone for you, or contact emergency services. If you or someone else is in immediate danger, contact your local emergency number right away, for example 112 in the EU, 911 in the United States and Canada, 999 in the United Kingdom, or 000 in Australia.

7. Not professional advice

KUN is entertainment. Characters do not give medical, mental health, legal, or financial advice, and KUN is not a substitute for therapy or for human support. Characters will not give financial advice or ask you for money.

8. Reporting a problem

You can report any AI message in the app. Long-press the message and choose Report, then pick a reason, such as sexual content, hate, violence, or self-harm. We review reports and act on them. A person may look at the message you reported. To raise a safety concern in any other way, email support@kun.chat.

9. What we count, and our annual report

Every time KUN shows a crisis card, it adds one to a counter kept by country. The counter holds numbers only: no message text and no user identifier. If you have allowed analytics, the app also records that a card was shown, without the text of your messages.

As California's SB 243 requires, KUN will report to the California Office of Suicide Prevention each year, beginning July 1, 2027, the number of times it showed a crisis service referral in the previous calendar year. The report contains counts only, and no personal information or message content.

10. Contact

For safety concerns or questions about this page, contact Anatoliy Belobrovik, Mejía Lequerica 34, Entresuelo 2A, 08028 Barcelona, Spain, at support@kun.chat.