📱 I built, tested, and deployed an entire web app to my private local network, from my phone.
Not “I wrote a prompt on my phone, then finished it properly on a laptop.”
The whole loop:
- described the problem
- reviewed the implementation
- watched the agent work and corrected it when needed
- ran the tests
- opened the app over my local network
- tested it from the same phone
- published the project
The result is Tabletime: a small app for answering the surprisingly difficult question, “What board game should we play tonight?” 🎲
You give it the number of players and the time available. It checks your own collection, including setup and cleanup time, and shows only the games that genuinely fit. You can also filter by complexity, competitive/co-op style, tags, and the age of the youngest player.
It imports public BoardGameGeek-sourced metadata, keeps the household library on the home computer, and works from both my wife’s phone and mine on the same Wi-Fi, the end goal all along.
And one detail I particularly like: the product itself contains no AI. It is a deterministic local app. AI helped build it; the user does not need to pay for model calls every time they want to pick a game.
Kudos to GPT-5.6 Luna on xhigh/max reasoning. It was fast, capable, and genuinely good company through the build. Also: Codex remote control is an absolute beast. 🔥
This was the first time my phone felt less like a time-waster and more like an actual tool for learning and building, work I would usually delegate to a PC.
I still had to supervise. I still stopped wrong turns, checked the details, ran the tests, and decided when the result was good enough.
But the place where that work could happen changed completely.
That feels more significant than “coding on a phone.” It feels like the development environment is starting to follow us, instead of the other way around.
Would you build a real project entirely from your phone?
More notes on engineering, AI and the systems around the work.
Back to all writing