Warmtype Join the Android test

If a chat mentions self-harm

What Warmtype does when a chat mentions self-harm or suicide, and where to find support.

Effective 16 October 2026.

If you need help now

If you or someone else is in danger right now, call your local emergency number.

These services offer free support. Your phone company's usual call, text or data charges may apply. Each one is for people in that country.

What Warmtype does

Warmtype's AI Reply suggests replies to a chat on your screen. When our check finds that a chat mentions self-harm or suicide, from you or from the other person, Warmtype does not write a reply, and the AI is not asked.

Instead, the app shows a short message with a free help line, or a link to find one. There is always a "Find a helpline" button (findahelpline.com). When we can name a line for where you are, there is also a button for that line. Older versions of the app show only "Invalid". Update the app to get the help message.

We pick the line from where your internet connection appears to be. This can be wrong, for example on a VPN or when you travel. If it is, use the "Find a helpline" button.

Our check looks for common phrases in English and in these languages: Arabic, Bengali, Bulgarian, Catalan, Chinese, Croatian, Czech, Danish, Dutch, Estonian, Finnish, French, German, Greek, Hebrew, Hindi (also typed in Latin letters), Hungarian, Icelandic, Indonesian, Irish, Italian, Japanese, Korean, Latvian, Lithuanian, Macedonian, Malay, Maltese, Norwegian, Polish, Portuguese, Romanian, Russian, Serbian (Cyrillic letters), Slovak, Slovenian, Spanish, Swahili, Swedish, Tagalog, Tamil and Turkish. It can miss things, and it can sometimes stop a harmless chat.

What we keep

We do not store the chat. When a chat is stopped, our server keeps a short note that a chat was stopped. The note holds no words from the chat, no country and no internet address.

Keeping young people safe

Anthropic, our AI provider, asks apps that people under 18 may use to say publicly how they keep them safe. This is what Warmtype does today:

  • Warmtype is for people aged 13 and over (see the Terms, section 1). If we learn that a user is under 13, we delete the server data we hold for that install (see the Privacy Policy).
  • If the other person in a chat says they are under 18, AI Reply does not write a reply, and the AI is not asked. The same happens when a chat mentions self-harm or suicide. Which languages the check understands is described above.
  • Our Terms forbid using Warmtype to create sexual content involving anyone under 18, or to harass, threaten, deceive or impersonate anyone.
  • The AI features are named as AI, and before your first AI use the app shows a notice naming the AI companies.
  • You can tap Report on any AI reply. We read reports, keep them for 90 days, and use them to make the app safer.

What Warmtype is not

Warmtype is a keyboard. It is not a counsellor or a crisis service. It cannot tell whether someone is safe, and it does not contact anyone for you.

Questions

Email support@warmtype.com.