Instruments, not oracles.

You make the call before you can see the answer. The target is hidden in code; the statement, the criterion, the impression are sealed and timestamped before the outcome exists. Reveal is one-way — you can't peek, then quietly discard a bad run.

An honest test is one you can't talk your way into passing after the fact. Everything else here depends on this one.

Every score is measured against a fair baseline — random chance, or your own prior — so a "hit" has to earn its keep. Stats stay hidden until there's enough data to mean something; below the line you see how much more you'd need, never a noisy chart dressed up as insight.

And the instrument will happily tell you "chance." A flat result is the instrument working, not you failing. We would rather show you an honest "no signal yet" than a flattering number you can't trust. That honesty is the product.

Every app runs entirely on your iPhone or iPad — no accounts, no analytics, no servers. What you record stays yours because there is nowhere else for it to go.

The App Privacy label reads "Data Not Collected" not as a promise but as the only honest answer: there's no server to collect it on. Where an app offers optional paid sync, it goes through your own private iCloud and nowhere else.

The category around us uses AI to generate the reading — type a fragment, receive a prophecy. We use it the opposite way. Where a Quiet Optics app uses on-device AI at all, it reads a call you already sealed and checks it against an outcome that already exists.

It can dock credit; it can't invent it — the displayed score is capped by your own hit/miss judgment. It never tells you what will happen next. And one app, Reticle, uses no AI at all, on purpose: your calibration curve is your data, not a model's guess about your data.