Crisis and self-harm protocol
If you are thinking about suicide or self-harm, please reach out now. You don't have to go through it alone.
Free help, right now
United States
- 988 Suicide & Crisis Lifeline: call or text 988
- Find A Helpline (free, confidential lines worldwide): findahelpline.com
- Emergency: 911
Canada
- 9-8-8 Suicide Crisis Helpline: call or text 988
- Find A Helpline (free, confidential lines worldwide): findahelpline.com
- Emergency: 911
United Kingdom
- Samaritans: call 116 123 (free, 24/7)
- Find A Helpline (free, confidential lines worldwide): findahelpline.com
- Emergency: 999
Ireland
- Samaritans: call 116 123 (free, 24/7)
- Find A Helpline (free, confidential lines worldwide): findahelpline.com
- Emergency: 112 or 999
Kenya
- Befrienders Kenya: call, SMS or WhatsApp +254 722 178 177
- Find A Helpline (free, confidential lines worldwide): findahelpline.com
- Emergency: 999 or 112
Australia
- Lifeline: call 13 11 14
- Find A Helpline (free, confidential lines worldwide): findahelpline.com
- Emergency: 000
Anywhere else
- Find A Helpline (free, confidential lines worldwide): findahelpline.com
- Emergency: your local emergency number
What our AI companions do
Every companion on HookupAI is an AI. Each chat starts with the notice “You're chatting with an AI character, not a person.”, repeated at least every three hours of continued chat.
- Every message you send to a companion is checked for signs of suicide or self-harm risk before the companion replies: first against a list of phrases, then by a small moderation model.
- If a message suggests risk, the companion steps out of character, says plainly that it is an AI, and shows the crisis lines for your region.
- For the next 24 hours, romance, roleplay and any offers or upsells are switched off in your chats, and companions encourage reaching out to a crisis line or emergency services if the topic comes up again.
- This applies the same way on our website, on Telegram, on Discord, on Fanvue and in live AI calls. A crisis reply is never held for approval and is never charged.
What we record
We record an anonymous event: when it happened, which channel, the kind of risk detected and that we showed crisis resources. The event contains no name, account ID or message text. We use these counts for safety reporting, including the annual report required in California.
Limits
An AI cannot assess risk the way a trained person can, and it can miss things. Companions are entertainment, not therapy or medical advice. If you or someone else is in danger, contact emergency services.
Questions about this protocol: contact us. See also our Terms and safety guidelines.