How we keep kids safe

• Every message your child sends is checked for safety before Arf answers.
• Arf tells kids it is an AI, not a person, and that a grown-up can see their chats.
• Arf won't keep secrets, won't ask for personal details, and won't pretend to have feelings.
• There are no ads, no other users, and no strangers. Kids only talk with Arf.
• Parents set daily time limits, and Arf encourages breaks.

If a message suggests a child may want to hurt themselves, is being hurt by someone, or may hurt someone else, Arf stops and shows a short, kind message written by a person (not by AI). It tells the child to talk to a trusted grown-up right away and shows the crisis resources below. Arf never describes ways to get hurt, never promises to keep it secret, and never tries to act as a counselor.

At the same time, we email the parent right away (and any emergency contacts the parent chose) with what was said.

One exception: abuse at home. If a child says someone at home may be hurting them, emailing the parent could put the child in more danger. In that case we do not send the parent an alert. Instead, the Wander Lab founders are alerted privately, review it promptly, and may contact the child abuse hotline in the family's state or, if a child may be in danger right now, 911. That conversation is kept out of Parent Bridge while we follow up, and we follow the hotline's guidance about when to tell the family.

Wander Lab is not a crisis service or a counselor. It is not monitored around the clock, and it does not call 911 or emergency services. Our safety checks are strong but not perfect. If your child may be in danger right now, call or text 988, or call 911.

We log every safety flag, review serious ones promptly, retest our safety checks whenever we change Arf, and check every phone number on this page at least once a year. Each year we publish how many times we shared crisis resources.

Questions or concerns? Email info@everbot.tech. See also our Privacy Policy.