
Used at home, self-hosted
FitTrack
A training and nutrition app with a rule-based coach that reads my recovery.
Lab
My job is helping organisations adopt AI and new technology. The best way I know to understand adoption is to go through it myself: decide what's worth building, write the requirements, build it with AI, and then live with the result every day.
These are tools I designed for my own household and built with Claude Code. I own the product, the architecture and the requirements, and I review the code. I don't write it by hand. They all run on a small server at home, and my family uses them.
Each one feeds back into my client work. People adopt what fits their routine, trust what they can understand, and drop what makes them feel behind. None of these apps calls an AI model. Where something looks smart, it runs on rules I designed, and each page explains how.

Used at home, self-hosted
A training and nutrition app with a rule-based coach that reads my recovery.

Running 24/7 on a mini PC at home
A private smart-home control panel that runs entirely on my own hardware.

In use at home by two parents
A shared baby log for two parents, synced live between our phones.

In use on our kitchen tablet
A kitchen app that knows what's in the cupboards and writes the shopping list for you.

In use at home by two people
A shared chore app that uses friendly competition, not guilt, to get things done.