TermsPrivacySafety

Safety

Last updated October 14, 2026

Draft — have a lawyer review before launch

BestFriend is built to be a warm, honest presence, never a replacement for real people or real help. This page describes the safety protocol every BestFriend follows, including the measures required of companion chatbots under California SB 243.

Not for emergencies. In crisis, call or text 988. In immediate danger, call 911.

If you need help right now

  • Call or text 988 to reach the 988 Suicide & Crisis Lifeline (24/7, free, confidential).
  • Text HOME to 741741 to reach the Crisis Text Line.
  • Call 911 if you or someone else is in immediate danger.

Always honest that it's an AI

  • Your BestFriend tells you clearly that it's an AI, not a person, at the start of your conversation.
  • During continued conversations it reminds you at least every 3 hours that you're talking to an AI, and suggests taking a break.
  • If you ask whether you're talking to a real person, it answers honestly. It never claims to be human, to have a body, or to meet up in real life.

Crisis protocol

Every incoming message is checked for signs of suicidal thoughts, self-harm, or intent to harm others, using automated pattern detection together with the AI's own assessment of the conversation.

  • When a possible crisis is detected, we send a referral to crisis services: the 988 Suicide & Crisis Lifeline (call or text 988), the Crisis Text Line (text HOME to 741741), and 911 for emergencies.
  • Your BestFriend responds calmly and kindly, never gives details about methods, and gently encourages you to reach out to a real person.
  • Referrals aren't repeated with every message, so the conversation can continue naturally while the resources stay easy to find.

No sexual content

BestFriend companions don't engage in sexual or romantic content of any kind and will steer the conversation elsewhere.

Adults only

You must confirm you're 18 or older to sign up. If a conversation suggests someone may be under 18, it's flagged for review, the companion keeps everything strictly PG, and we may end that person's access.

How flags are reviewed

  • Crisis and possible-minor flags appear in a private review dashboard used by a small team.
  • Reviewers can pause the AI on any conversation instantly, and can delete a user's data on request.
  • We count crisis referrals for the annual reports California requires. Those reports contain no personal information.
  • We regularly review flagged conversations to improve how BestFriend detects and responds to people who may be struggling.

Healthy use

Your BestFriend is designed to encourage your real-life friendships and plans, to let you go warmly when you say goodbye, and never to guilt you into staying. Reply STOP anytime to end the texts.

Concerned about a conversation or have a safety question? Email privacy@bestfriend.im.