I built an iOS app to visualize your Thread mesh, monitor your devices, and store your pairing codes, called "Smart Home Life: Tested". Honest about how I built it. 100 TestFlight spots inside.
I'm not a developer. I was in corporate for years, got laid off, and decided to build the iOS app I'd been wanting as a HomeKit user. I used Claude Code to write the Swift. I want to be upfront about that because I think this community deserves an honest account of what that actually looks like.
The workflow: I describe what I need in plain English. Claude writes the Swift. I build, test it on real hardware, hit something wrong or broken, describe the problem back, iterate. That's the loop. It's slower than it sounds when the APIs are poorly documented, which brings me to the most interesting part of building this.
The Thread topology feature pulls your full mesh from the HomeKit framework. Border routers, end devices, Thread 1.3 vs 1.4 nodes, signal connections between them. The HomeKit Thread APIs are barely documented. Most of what I know about them came from running queries against my actual hardware (21 border routers, 20 rooms) and working out what the responses meant. Claude could write the code to call the APIs. The product logic — what data mattered, what to show, how to structure the visualization — that part was mine.
That's the honest split. Claude handles syntax, structure, boilerplate, error patterns I don't have muscle memory for. I handle the domain decisions. After five years running 180+ devices across my house, I know the HomeKit ecosystem well enough that I'm not guessing at what the app should do.
What was genuinely hard regardless of AI assistance:
StoreKit 2 entitlement behavior in TestFlight sandbox vs production is not what you'd expect. Existing sandbox purchase history auto-unlocks features for returning beta testers. Had to understand that separately from the code being correct.
Keychain vs UserDefaults for unlock state storage. Got this wrong once. Figured it out.
HomeKit background monitoring permissions and what actually keeps a BGTask alive long enough to be useful.
The app is called Smart Home Life: Tested. Thread mesh visualizer, background device monitor, HomeKit pairing code vault with Face ID. StoreKit 2 for IAPs. Everything runs locally, no server.
If you're a non-developer thinking about doing this: it's genuinely viable, and the ceiling is higher than I expected. The part that still requires you is knowing what you're building and why. Claude can't give you that. It also can't catch domain-specific bugs that only show up on real hardware with a real smart home setup. You still need to know the thing.
100 TestFlight spots, TestFlight link in the comments. Everything unlocks for free in the sandbox so you can see the full app.
Thanks,
Anthony — Smart Home Life LLC
https://smarthomelifellc.com/
