Preflight

How it works

Three things decide whether a check is worth anything: what it looks at, where its rule came from, and whether it tells you when it could not look.

1. Your browser measures. Our server judges.

When you add a screenshot, your browser reads its dimensions, file size and colour type on your own machine, and sends us those numbers. The image never leaves your device.

That is two benefits from one decision. Your unreleased screens stay yours, because most tools in this space upload your files to a server and then compete on how fast they delete them afterwards. And our rules stay on our server, where they cannot be read out of the page.

Your listing text is sent to us. We have to read the actual words to check them, and that text is going to be public on the store in any case. We would rather draw the line precisely than claim a tidier one.

2. Play is measured. Apple is documented.

This distinction decides what we charge, so it gets its own section rather than a footnote.

The Play checks come from real submissions. We have shipped to Google Play and been corrected by it. Some of what we check is in the documentation; the more useful part is not:

The Apple checks come from Apple’s published guidelines. We have not submitted an app to Apple, so we have measured nothing there. The limits are correct and the checks are real, but they carry documentation authority rather than experience, and that is a genuine difference in what a check is worth.

So Apple is free, and every Apple report says on its face where its checks came from. It becomes paid when it is earned, and not before. If Apple rejects you for something we missed, tell us. That is how those checks stop being documentation.

3. A check that could not run says so

If you do not give us an icon, the report says the icon was not checked. It never says the icon is fine.

That sounds obvious, and it is the thing most tools get wrong, because an empty result and a clean result look identical on screen. Every report carries a list of what could not be checked, and that list is as much a part of the answer as the findings.

What we do not check

We do not open your build: no binary, no bundle, no code. We check your listing, and the declarations around it: the privacy policy that has to load, the account deletion that has to exist, the demo credentials a reviewer needs to get past a login wall.

A clean report is not a guarantee of approval. Review includes human judgement we cannot see. We tell you what we can check, and we are precise about the rest.