Built with AI, run by a runner.
What I do
I decide what gets built and why. Every change is one I asked for or agreed to. Every change is built and tested automatically before it reaches anyone. I use each build on my own phone, for my own training, and nothing goes to testers or the App Store until I decide it’s ready.
Tested on real running, every day
I use Mile32 every day for my own training, so every build gets tested on real runs, not just on a test device. Since I started building it in April 2026 I’ve logged over 1,000 km across more than 180 runs in it, and many of the fixes in the changelog began as something I noticed on one of them.
What every change goes through
- A build with zero warnings. A change that leaves a compiler warning is not finished.
- Over 300 automated tests of the app’s calculations, from training load to plan dates to unit conversions, run before a change is kept.
- A second build and test on Apple’s Xcode Cloud for every change, so nothing relies on one computer.
- A look at the screen. A change you can see is checked on a running copy of the app wherever it can be, and on my own phone before it ships, not just compiled.
- A written record. Every change is logged with what it fixed and why, and the changelog lists what has shipped.
What AI doesn’t touch
The app contains no AI and calls no AI service. Your runs, health data, plan and settings stay on your iPhone and in your own iCloud, exactly as the privacy policy describes. I can’t see them, so they are never used to build or test the app.
Why say so
You’re trusting Mile32 with your training and your health data, so you should know how it’s made. Using AI lets one person build and keep improving an app this size. The decisions, and the responsibility for getting them right, stay with me.
If something’s wrong
Tell me. Settings › Send Feedback in the app, or support@mile32.app, reaches me directly.
Last updated 18 November 2026.