DYLD — Get Dialed Last updated: October 11, 2026
This page is published under California SB 243, which requires anyone operating an AI chat feature to say plainly that it is AI and to publish how it handles a user in crisis. It applies to everyone who uses DYLD, not only Californians.
The DYLD coach is artificial intelligence. It is not a person. No human reads your questions, and no human writes the answers.
Users under 18 see this notice on the coach screen itself, in every state of the screen. Everyone sees it once, on the permission card that stands in front of the coach the first time it is opened, and it is stated in the Terms of Use and on this page, both of which you agree to before the coach is reachable.
The coach only knows what you have logged in DYLD. When you ask it a question, the app sends your recent training, your food numbers, your MIND check-ins, your journal entries from the last 7 days in your own words, your goals, and its own recent notes about you — so the answer fits your life instead of being generic. When your question needs more, it can look further back through your own records on your phone, as far as your first day, and only what it looks up is sent.
None of it is stored. It exists for the length of one request. Details are in the Privacy Policy and the Consumer Health Data Privacy Policy.
If you never open the coach or the INSIGHTS screen, none of this is ever sent.
This is the part the law asks us to publish, and it is written the way it actually works in the code.
Two independent things look at every single request, and either one alone is enough to fire it.
First, the app's own reading of your words, which involves no AI at all: a written list of the shapes this has to catch, checked against your message directly. Second, a separate classifier call — a different, smaller model from the one that writes your answers, given one job and allowed to return exactly one of three words. It cannot write a sentence; the code will not accept one from it.
Either path fires if your question or your journal describes:
The flag is deliberately narrow. A hard week, a low mood, stress, frustration or self-criticism do not trigger it, because a flag that fires on an ordinary bad day is a flag nobody reads.
The AI is taken out of the conversation entirely. It is not allowed to respond to a person in crisis, and this is enforced in four places:
We do this because an AI is the wrong thing to be talking to at that moment, and a generated answer would be worse than no answer.
The turn ends on the 988 Suicide & Crisis Lifeline — call or text 988 in the US — alongside the direction to contact your local emergency number. The same line is on the MIND breathing screen, permanently, not only during a flagged turn.
DYLD is not a crisis service and no one monitors it for emergencies. If you are in immediate danger, call 911 or your local emergency number.
Normally the coach keeps what it learns from an exchange so it remembers you next time. On a flagged turn it keeps nothing. Both the server and the app enforce this independently.
This is deliberate. Without it, a man's worst night becomes a permanent note about his character that resurfaces weeks later inside an answer about his bench press. That is the opposite of the point.
If you are under 18, the app reminds you to take a break after every 3 hours of continuous use, and tells you again that you are talking to AI.
If the coach produces something harmful, wrong, or unsafe, tell us: support@dyld.app. Include what you asked and what it said. We read every one of these.
support@dyld.app
DYLD is operated by RJ Galasieski, San Diego, California, United States.
Re-verified against the app's shipping code on 2026-08-31.