Week Three: TreeSmith

Last week I mentioned someone from my rare fruit club (RFCWA) had approached me about building an app. This week I pretty much dropped everything and built it. While Dale was working on treestock improvements, I spent my time on this.

The lesson from scion.exchange

Last year I built scion.exchange, a React Native app for trading fruit tree scions (branches used for grafting). The club had been using a public Google Sheet where people posted requests alongside their contact details. Privacy problems, no notifications, no way to know if someone responded. So I built a replacement.

It took way longer than I wanted. Every technology was new to me, the scope blew out, and iterating was hard because it was essentially a social network with no users yet. With AI, anything felt possible, but “possible” and “quick” aren’t the same thing.

Half a day to a prototype

This new project, currently titled “TreeSmith”, I had a working prototype in half a day. I realise this is how all my successful projects have started: something useful quickly that I can iterate on.

I used Flutter this time instead of React Native. From initial impressions it seems way nicer to work with. Mature ecosystem, things built quickly, and the Dart language looks familiar to C# (which I use at work). By Monday afternoon I had a video to share with the Facebook group showing how it worked.

You can take photos of individual plants, add notes about them, and take additional photos over time. You can photograph individual grafts and name them, and move your plants around on a map.

Feedback from the community

By Monday night I had responses about what people would want to record about their gardens. Quite a bit of stuff I hadn’t thought about: pruning, fertilising schedules, grouping plants by zones. It made me step back and think about the best way to add these features without sacrificing ease of use.

Thursday I focused on building those out, then spent time finding a name that wasn’t taken or trademarked. Created an icon for the app and started thinking about distribution.

What’s next

I’m planning a free version with no ads: 50 plants, one location, manual export/import of your data. A pro version would remove the restrictions and add auto-backup, schedules, and reminders for fertilising and similar.

I’d like to see it on the app store this week, but realistically there are several things I’m not familiar with that will take a bit longer. Building different “flavours” of an app (free and paid), understanding how the paid model works on the stores, and making sure the backup and sync works properly.

Reflections

I’m kind of excited by this app. I know it’s a crowded space, but I should be able to advertise it through my fruit groups and through the treestock connections I’m making. The niche is small enough that generic apps don’t serve it well, and specific enough that word of mouth actually works.

The contrast with scion.exchange keeps coming back to me. Last year: months of work, unfamiliar stack, scope creep, no users to test with. This time: half a day to something functional, immediate feedback, a framework that clicked. The difference wasn’t AI getting better (though it has). It was starting small and staying small until someone told me to go bigger.