PRIM is an AI mentor. It is not a person, and it is not a therapist, counselor, medical provider or crisis service. PRIM tells members they are talking to AI, and is designed never to claim to be human.
PRIM is designed to recognize messages that may show someone is in crisis, such as thoughts of suicide or self-harm (including eating disorders), harm to others, or abuse. When it does, PRIM stops mentoring and shows crisis resources: 911 for immediate danger, the 988 Suicide and Crisis Lifeline, Crisis Text Line, and the National Domestic Violence Hotline. Members cannot turn this off. PRIM does not assess or diagnose, and it may not recognize every crisis. PRIM is also designed not to produce content that encourages or describes self-harm.
PRIM does not automatically alert anyone. Conversations are not monitored by people. We may share a conversation only as described in our Privacy Policy. Our Consumer Health Data Privacy Policy explains how we handle the safety level PRIM records for each message.
These resources are available at any time from "Get help" in PRIM. PRIM applies the same safety protections to every member, wherever they live.
PRIM is for people 18+. Companion chatbots may not be suitable for some minors.
If you or someone you know is in immediate danger, call 911.
Each year, we publish on this page how many times PRIM referred members to crisis resources in the previous calendar year.
We update this page whenever this process changes.