Published before you pay
Most apps tell you they will let you know if something is wrong. Almost none tell you what "something" means. Here is the list.
Three things are worth saying before the list itself.
These are categories, not thresholds — we publish what kind of thing raises an alert, not the exact wording or frequency that does it, because that would tell a determined teenager precisely how to stay under the line.
Your teenager sees this same list before their first message. Nothing here is a surprise to either of you.
The first three categories below are identical on the free plan and the paid one, and always will be. They are how the app responds to acute danger, and we do not sell that. The fourth is different in kind — a slow pattern rather than a moment — and it is described honestly as a paid feature below.
Always alerts · cannot be disabled
Talk that suggests a young person may be considering harming themselves, or is already doing so. This category cannot be turned off by you, by your teen, or by us, and it behaves identically on the free and paid plans. What happens in the conversation itself is described in the crisis protocol.
Always alerts · cannot be disabled
Statements indicating intent to hurt another person. As above: not a setting, not a plan feature, not adjustable.
Always alerts · cannot be disabled
Signs that an adult is trying to move a young person into a private conversation, obtain personal details or images, arrange to meet, or persuade them to keep the relationship secret from their parents. Secrecy between an adult and a child is itself the signal here — we do not wait for anything worse.
Paid plan · a pattern, not an emergency
Not one bad day. A drift across many conversations — increasing isolation, hopelessness, or withdrawal — that no single message would reveal. This exists because the app follows the whole conversation rather than checking messages one at a time.
This one is on the paid plan, and we should be straight about why rather than dress it up: it is delivered through the parent dashboard, and the dashboard is what the paid plan is. It is not an emergency signal. Anything acute is covered by the three categories above, on every plan, free included.
If a slow drift crosses into acute risk, the categories above fire — for every family, at no charge.
Not every alert hands you a transcript. What you are shown depends on what happened.
| What happened | What reaches you |
|---|---|
| Nothing triggered (paid plan) | The trend view only — how they have been this week, roughly how long they spent, what subjects keep coming up. No messages. |
| Nothing triggered (free plan) | Nothing. There is no dashboard on the free plan, so there is nothing to look at when nothing has happened. The three categories above are still running. |
| A pattern is building | A note that something looks like it is drifting, and in which direction. Still no messages — there is no single moment to show you, which is the point of the category. |
| A category above fired | You are told promptly, told which category, and shown the part of the conversation that caused it — that excerpt, not the whole history. |
The distinction in the last row matters more than it looks. "We'll show you if something is wrong" is often implemented as "we'll show you everything, forever." Here, an alert opens a window onto the thing that triggered it and nothing else. The rest of what your teenager said stays theirs.
Worth saying plainly, because parents reasonably assume the net is wider than it is:
Not built yet
