I Built My Wife an App. I Still Can't Code.
Last Saturday I built an app. A real one. It's on the internet right now. A real person uses it every day.
That very real person is my wife, Fran, but hey, gotta start somewhere, right? ;-)
It took about 7 or 8 hours and I don't even know how to code.
I need to back up for a second, but only a second.
Fran is doing a 100-day painting challenge where, as is possibly evident by the name, she’ll paint something every day for 100 days, share each day publicly and track the streak. She was managing the whole thing manually. Posting to Instagram, keeping count in her head. And I thought: what if she had her own app for this? Something where she could log each session, see her progress climbing a visual mountain toward Day 100, and share a link with anyone who wanted to follow along.
It ended up being more a cool thing to show her friends and family and we highlighted her IG profile so she got some followers (maybe even broke 500?)
So I built it. It's called Fran's Peak. (The AI named it. Neither of us are fans. But it's V1, whatever, we moved on.) An important note: despite being the CEO of a software company, I’m non-technical. I spent about 60 days learning to code around 5 years after we started the business. That’s it.
So how did I build an app?
I'd been watching friends in Lisbon use an AI coding tool called Claude Code. My friend Noel showed me a personal motivation app he'd built for himself, saved it right to his phone's home screen, looked and felt like a real app, no App Store needed.
Now I’m used to seeing stuff like this because many of my friends are technical. But my friends building things this week weren't engineers. They were people like me. That broke my brain a little about what finally might be possible (for me).
So Saturday morning, I opened Claude Code and told it what I wanted to build for Fran. At the advice of friends I asked the AI to interview *me*. To ask every question it needed to get clarity on what I actually wanted. It asked maybe 20 questions. What's the challenge? How does she track progress? What should the morning ritual feel like? What's the vibe?
Then I asked it to take all my answers and create three documents: product requirements, technical requirements, and design guidelines. I went through each one with a fine-tooth comb. Gave feedback. Iterated. And once they felt right, I basically said "go build this."
And Claude Code did.
I chose to build with Rails and React, the same stuff ZenMaid runs on. This gave me an added benefit in that while I was learning to build, I was also starting to understand our own codebase a bit. Two birds, one stone.
As the AI built, I asked it to explain what it was doing. Not all of it stuck. A lot flew over my head. When it told me to type commands into the terminal, I'd look up what they actually meant. Really basic stuff. Stuff an actual developer would find embarrassing. But here's the beautiful thing about AI, no question is too stupid. I had the same concept explained to me three, four, five different ways until it finally clicked. And some of it didn’t click, and that’s ok. Progress is progress!
Somewhere in there, time disappeared. The kind of flow state I haven't felt since the early days of ZenMaid, when everything was new and every small win felt enormous. I was tucking little surprises into the app, easter eggs for our pup Lola, custom milestone messages, little delights because I knew she'd be the one opening it every morning and I wanted her to smile. It’s fun to think about all these details while creating something :-)
Saturday night, it worked on my computer. Sunday morning, I got it live on the internet with a real URL. Fran started using it. I sat there staring at my laptop like an idiot thinking *holy shit, I actually made this.*
I should be honest about the AI's role here because I don't want this to read like a sponsored post. I'm not a developer now. I know I'm not. The best way I can describe working with Claude Code is like having a really talented but slightly overeager junior dev team sitting in the room with me.
They're fast, they're brilliant, they'll build whatever you ask. But they might cut corners or make decisions that a seasoned engineer would side-eye. You have to check their work. You have to stay engaged. You have to think.
Which is kind of the perfect setup for someone trying to learn.
I'm already planning to rebuild the whole thing from scratch, not because it's broken, Fran uses it every day but because I made enough mistakes along the way that I want to run it back with everything I know now. The app was the output. Education was the point.
But here's the part I can't stop thinking about. Four days. That's how long it took me to go from "I barely know what a terminal is" to "there's a live app on the internet." Most of that was a single Saturday afternoon. Me, a guy whose only code contribution to his company is the year changing automatically on Jan 1 each year (yes I’m proud of this), built a working app between lunch and dinner (a very late dinner).
I don't know exactly what comes next. Actually, that's not true. I'm rebuilding this for multiple users so Fran's artist friends can join. But for the first time in a long time, the list of things I *could* build feels way longer than the list of reasons I can't.