Apple said no, and it was the best thing that happened to tab.

tab. is live now. But it didn't walk in the front door. The first time I submitted, Apple sent it back.

Two problems. How I handled Sign in with Apple, and how I linked out around subscriptions. And I'll be honest with you about the first feeling, because everyone has it and nobody admits it: I wanted to argue. I'd built the thing. It worked. My TestFlight people liked it. And some reviewer I'll never meet just told me no.

Here's what flipped it for me. I stopped reading the rejection like an insult and started reading it like a spec. Because that's what it is. Apple almost always enforces exactly what it wrote down — the guideline isn't a mood, it's a rule with a number on it. Once I read it that way, the fix stopped being a fight and became a checklist.

The Sign in with Apple thing? It was never "add the button." It's that Apple's login can't be a second-class option sitting next to the others. The subscription thing was the same shape — it's about not routing people around Apple's payment system in the ways the rules say you can't. Boring when you say it plainly. That's the point. It was always boring and specific, I just wanted it to be unfair.

Did the AI save me here? Not in the way you'd think. It didn't see the rejection coming. Where it earned its keep was the turnaround — rewriting the auth flow and adjusting the subscription surface took an evening instead of a week. That's the whole difference. When the fix is cheap, a rejection is a speed bump. When the fix is expensive, a rejection is a funeral.

Approved in June. Live ever since. And now I treat the guidelines like a test suite that runs last and can't be faked — I check auth, payments, anything that links out, before I ever hit submit. The second submission is always better than the first. The trick is just making the first one cheap enough that you don't take it personally.