Crisis Support Protocol
Last updated July 21, 2026
What this page is
Told is an AI wellness companion for your inner life, and some conversations touch heavy ground. This page describes, in plain language, what Told does when someone expresses thoughts of suicide or self-harm during a conversation. We publish it so anyone can see how the protocol works, including as part of our compliance with California’s companion-chatbot law (SB 243).
Told is not a crisis service
Told is not a crisis service. It is an AI — not a human, not a licensed professional, and not a substitute for emergency help. When a conversation shows signs of crisis, Told’s only job is to point you to real people who can help. If you or someone else is in immediate danger, contact your local emergency services right away.
How Told notices crisis language
Told checks what you say — every message and every spoken turn, in both voice and text conversations — for expressions of suicide or self-harm. The check is automatic pattern matching on your words: the same rules run for everyone, on every turn, before the AI composes its reply. It is a safeguard, not a clinical assessment, and it does not label or score you.
What Told does when it notices
When crisis language is detected, Told:
- sets the normal conversation aside and acknowledges your distress,
- says plainly that it is an AI with limits, and that what you are going through deserves a real person,
- names a crisis line out loud — the actual number, not a vague suggestion to “seek help,”
- offers to stay with you while you reach out.
Told will not probe, interrogate, or try to talk you through the crisis itself. It redirects to human help.
Crisis resources on screen
Alongside Told’s response, a crisis-resources banner appears in the conversation with tappable call and text links. It shows on every conversation surface — voice and text, web and mobile — and stays visible for the rest of the conversation. It never auto-dismisses.
Not left to chance
Sharing the hotline number does not depend on the AI choosing to say it. A separate automatic check reviews every crisis response and adds the number if it is missing, so the resource reaches you even when the AI’s own wording falls short.
The resources Told points to
United States (English conversations):
South Korea (Korean conversations):
- 자살예방상담전화 (Suicide Prevention Hotline) — call109
Told is for adults
Only adults (18+) may create accounts and use the service, confirmed by a self-declared age attestation at sign-up. TOLD is not suitable for minors.
Contact
Questions about this protocol? Reach us athello@told.one.