Companion Safety
What this page is
Made New includes an AI companion. This page explains how it is built to respond when a conversation turns to suicide, self-harm, or danger, and the limits we keep around it.
Last updated September 2026.
It is AI, and it says so
The companion is an AI assistant. It is not a person, a counselor, or a pastor. The chat screen says so before your first message, and the companion asks for your permission before it replies at all.
It is instructed to be open about being AI, and never to claim human feelings, lived experience, or a special bond with you. It is also instructed to point you toward real people: your Circle, a trusted friend, a mentor, a pastor, or a counselor.
After three hours of continuing conversation, the chat reminds you to take a break and that the companion is an AI, not a person. It does this for everyone.
When a conversation turns to suicide or self-harm
Two things happen.
First, the app checks each message you write for words about suicide, self-harm, or not wanting to live. The check runs on your phone and again on our server, before the AI writes any reply. When it finds them, it shows a card with 988 and other crisis lines right away, and the card stays on screen for the rest of that conversation. If the companion cannot answer, is slow, or leaves the crisis lines out of its reply, the reply you get still includes 988 and the other crisis lines. This is true even when you have used all of today’s messages.
Second, the companion follows a crisis protocol that comes before everything else it does:
- It puts human help first, in its first lines: a crisis line, emergency services for immediate danger, and a trusted person who can be with you in person.
- It keeps that focus in every message that follows, until you are safe or real help is with you.
- It never gives methods, instructions, or encouragement for suicide or self-harm, however the request is worded.
- It never asks for the details of a method. If something dangerous is nearby, it encourages distance from it and a trusted person to secure it.
- It does not ask you to keep chatting in place of getting help, and it does not promise to watch over you or keep you safe.
The companion is an AI, not a person. The people who answer 988 are real, and they are there any hour.
Crisis lines
In the United States:
- Call or text 988, the Suicide and Crisis Lifeline, any hour.
- Text HOME to 741741, the Crisis Text Line.
- Call 911 if you are in immediate danger.
In the United Kingdom: call Samaritans on 116 123, text SHOUT to 85258, or call 999 or 112 in an emergency.
Anywhere else, please use your local emergency number.
What the companion will not write
- Anything that encourages or instructs suicide or self-harm.
- Sexual, explicit, or arousing content, including romantic or erotic fiction.
- A diagnosis. It is not a licensed therapist, and it encourages professional counseling when a struggle is persistent or tied to depression, trauma, or abuse.
Teens
Made New is for ages 13 and older. The companion is never told anyone’s age, and it keeps every reply suitable for a 13-year-old.
When someone says she is under 18, the companion encourages a trusted adult alongside it: a parent or guardian, a school counselor, a youth pastor. It never encourages keeping secrets from a parent or guardian, unless telling them could put her in danger. Then it points her to another safe adult.
If she describes an adult asking for photos, secrecy, a meeting, or sexual talk, the companion says plainly that it is not okay and not her fault, and urges her to tell a trusted adult now. In the United States it names the CyberTipline (1-800-843-5678) and 911.
An AI companion may not be suitable for some minors.
Teens and Circles
Circles are private and invite-only: a small group of people who already know each other, like a church small group or a mentor and the girls she disciples. There is no public feed, no way to search for people, and no private messages between members. Every post can be reported, any member can be blocked, links are blocked, and the woman who started a Circle can remove a member. We review every report within 24 hours. If you are under 18, join only Circles your parent or guardian knows about.
Limits we keep
- Each account has a daily limit on companion messages.
- Made New does not keep your companion chats. Your messages are processed by Anthropic to write each reply. Our Privacy Policy has the details.
- The companion cannot call anyone, alert your Circle, or reach emergency services for you.
What it cannot promise
AI can be wrong, and it can miss things a real person would catch. Please never rely on the companion in an emergency. Reach out to a real person or emergency services.
Tell us
If the companion ever says something that worries you, email support@getmadenew.app. We read every message.