How this was made

Last updated 15 August 2026

Why say so

Because every commit in the public repo is already tagged with it, so this was never really a choice between disclosing and not — only between saying it myself or letting someone else notice first. I'd rather say it myself, plainly, and let the app speak for whether it was worth doing this way.

What "AI-assisted" actually means here

I'm not a professional developer. I don't write Swift by hand. What I bring is a few years of wanting exactly this tool badly enough to spend a summer on it, and the judgment to know when something generated for me is wrong. In practice that's meant catching things the AI got wrong before they shipped: an email typo that would have sent tester feedback to a stranger's inbox, a demo screen that quietly stopped showing real tracker progress, a sync bug that would have silently dropped data between devices. None of that is a story about AI writing perfect code. It's a story about someone reviewing it like the mistakes were going to be theirs to live with — because they were.

Every product decision — what LevelSelect tracks, what it refuses to collect, what a run is versus a session, which games get a bespoke tracker — is mine. The privacy stance on the Privacy page isn't boilerplate; it's the same standard applied to how the app itself got built.

If you want to look

The repository is public, and it stays that way: github.com/timultuoustimes/levelselect. Every commit's history, including the AI co-authorship tags, is right there.