What we can and can’t see

Can’t see: your rules, agreements, limits, signals, messages, the names you use for each other, your photos. Any of it. It is encrypted on your device with keys we never hold, so it reaches us as noise.

Can see: that your account exists, that an encrypted record was written or changed and when, which collection it belongs to (the word “signals”, not the signal), roughly how much you store, and the push subscription needed to send a notification that itself carries no content.

So, to be blunt about signals: we can tell that something happened in your safety board at 11:40pm. We cannot tell whether it was green or a full stop, and we cannot read a word you typed with it. If that distinction matters to you, it should, and now you know exactly where the line sits.

No AI, anywhere: nothing you write is sent to a model, scored, summarised, or used to train anything. We could not do it if we wanted to, because we do not have your keys.

The one exception: beta feedback goes to us readable, and says so right on the form.