I Tried to Build a Trello Clone Twice. The Second Time Took an Afternoon.

What changed wasn’t me. What changed was the tools.

The first time I tried to build a Trello clone, it was 2021 and I gave up after three weekends.

The drag-and-drop kept fighting me. Cards would snap to the wrong column. Reordering broke on Android but not iOS. The list of “just one more thing” tasks kept getting longer while the app got worse. I closed the repo and told myself Kanban was harder than it looked.

Trello sold to Atlassian for $425 million. Fifteen years of polish on one very good idea — digital sticky notes. It always struck me as the kind of product a solo builder should be able to recreate. Not compete with Atlassian at scale, but at least make a version tuned for one niche they ignore.

The second time I tried, it was 2026. And it took an afternoon.

What changed wasn’t me. What changed was the tools.

I wasn’t writing gesture handlers anymore. I wasn’t tuning Reanimated worklets. I wasn’t debugging why cards on Android had a 40ms input delay. I opened an AI mobile app builder, typed a paragraph describing what I wanted, and had a working prototype on my phone in under an hour.

The paragraph was almost embarrassingly simple:

Build a project management app for small teams. Users see a list of boards on the home screen. Each board has three columns — Backlog, In Progress, Done — with cards that can be dragged between columns. Cards have a title, description, due date, and assignee avatar.

That’s it. That’s the whole spec. The tool — RapidNative, which generates React Native and Expo code from prompts — gave me the four core screens: board list, Kanban board, card detail, and new-card form. Everything navigable. Everything styled. Drag and drop working.

The first version wasn’t perfect. Nothing ever is. The column headers looked too heavy. The floating action button was in a weird spot. The card animations were fine but not delightful. So I did what I never could do in 2021: I clicked directly on the thing that annoyed me and typed what I wanted instead.

“Make the column headers pill-shaped with a subtle color per column.”

Done.

“When I tap the FAB, open a bottom sheet instead of navigating to a new screen.”

Done.

“Add haptic feedback when a card is dropped into a new column.”

Done.

Each change took seconds. I made maybe forty of them over the afternoon. Some worked immediately, some I had to re-prompt. But I never once had to think about the underlying code — just about how the app should feel.

Here’s the part I didn’t expect. The generated app wasn’t a mockup or a prototype in the disposable sense. It was real Expo code. I exported it, opened the folder in VS Code, and it was a normal React Native project. TypeScript. Clean file structure. The kind of code I’d have paid a freelancer a week to produce.

That mattered because it meant I hadn’t traded ownership for speed. If the app ever needs a custom backend integration — some proprietary API, a real-time sync layer I want to hand-tune — a human engineer can pick up where the AI left off. No proprietary runtime. No lock-in.

The traditional path to build a Trello-style app in 2026 still exists and still works. It’s what I would have done: React Native, Expo, Supabase for auth and real-time sync, react-native-gesture-handler and Reanimated for the drag-and-drop, Expo Notifications for push. That path takes a designer and two engineers three to six months and $40,000 to $120,000 — and it's worth it if your app has deep custom logic.

But most project management apps don’t have deep custom logic. Most of them are Trello with a niche twist. And for that class of product, the traditional path is now the wrong path.

The thing nobody warned me about — the reason my 2021 attempt failed — was that Kanban is deceptively hard to get right on mobile. Drag-and-drop on a touchscreen is a physics problem, not a UI problem. Card position, list scroll, screen-edge auto-scroll, haptic timing, cancel gestures. Every one of them interacts with the others. Every one of them behaves subtly differently on iOS and Android.

Getting that right takes weeks. It’s the kind of work where you feel like you’re not making progress even when you are.

The reason 2026 is different isn’t that AI writes prettier code than I do. It’s that AI has already done that work — and hands you the version that already feels right.

You spend your afternoon on the parts that make your app different, not on rebuilding the parts every Kanban app shares.

That’s the shift. It’s less “AI replaces developers” and more “AI does the boring parts so humans can do the interesting ones.”

I’m not going to tell you the app I built is going to displace Trello. It isn’t. But it doesn’t have to. It just has to serve one small group of users better than Trello does — freelance photographers tracking client shoots, in my case — and it’s already better because Trello has to be generic and mine doesn’t.

That’s the shape of most opportunities in software now. Not to beat the incumbent at their own game, but to build the version they’d never bother building, for the customers they’d never bother serving.

Fifteen years ago, the barrier to doing that was time and money. In 2026, it isn’t. If you’ve ever thought about building your own version of a tool you use every day, the tooling has caught up to your ambition. What’s left is picking a niche and getting started.

If you’re curious about the tool, it’s *here. Free to start, no credit card — you’ll have your first Kanban board running on your phone before your next coffee.*

If this resonated, a clap or two helps it find other builders. And I’d like to hear: what’s the tool you use every day that you’d rebuild for your own niche?