A times tables app for my nine-year-old

What it is

A times tables practice app for my daughter, running from her iPad home screen. She uses it regularly, which is the test that matters.

Why I built it

Partly a real problem, partly an experiment. I wanted to make practice feel less like homework, and to understand how far current AI-assisted tools could take a non-developer from an idea to something genuinely useful.

How I built it

Claude Code inside VS Code, published on Netlify, backed up in GitHub, with a shortcut on her home screen. Once Claude is running I can use Wispr Flow and just say what I want, which turns out to matter more than it sounds: the friction of typing was what had stopped me starting before.

Two decisions did the work. The first was three modes rather than one, because she is tested three different ways at school. Easy is multiple choice, hard makes her type the answer, and test mode gives her six seconds a question to mimic the real thing.

The second was the reward. Harder tables are worth more, hard mode beats easy mode, and points unlock bronze, silver and gold. What unlocks is the part that mattered: YouTube videos we chose together, not rewards I picked for her.

What didn’t work

There are still a couple of bugs, and my daughter finds them faster than I do. Generation is quick; quality control and edge cases still need attention.

What I took from it

Two things.

The barrier is now much lower than I expected. The gap between having an idea and having a working thing on a device is small enough that not building something has become a choice rather than a constraint.

The harder part was product judgement, not code. The mechanics were straightforward; understanding what she actually valued was what made the app useful.

She keeps asking when we can improve it, and she wants her six-year-old brother involved, which means a separate rewards profile full of Linkin Park videos.