I Built a LEGO inventory app

I Built a LEGO App and Shipped It — Just Because I Wanted To I built a LEGO app and shipped it. Not because there was a market gap — because I had the problem, and I wanted to see how far one person and AI could take it.

It started with two itches.

The first was a literal mess. I had LEGO everywhere — bins, drawers, bags on bags — and no real idea what I owned, let alone what it was worth. I'd buy stuff I already had. I'd sit on parts worth real money without knowing it. The collection had outgrown my ability to keep it in my head.

I also sell excess LEGO on BrickLink as a side hustle. If you've used BrickLink's upload flow… you know. Powerful, but clunky. Finding a part, setting condition, location, quantity, and listing it took way more clicks than it should — so I'd put it off, and the pile just grew.

The second itch was professional curiosity. I come from a product management background, and I genuinely wanted to know: in 2026, how hard is it for one person to build and publish a real, store-ready app with AI doing the heavy lifting? I wanted a meaty, end-to-end project to actually push AI and modern tooling on — not a toy demo, but something I'd have to ship, support, and put my name on.

So I did the thing product people do instead of cleaning their room: I built the app. It's called BrickStack.

What it does Point your camera at a brick and it recognises the part — no barcodes, no squinting at tiny mould numbers, just visual identification. Or search any part or set (Rebrickable API) and add it in seconds — batch mode locks fields like location and condition so I can rip through a whole drawer fast. See what it's worth — check the latest market prices and set your own, so you finally know the value of what you're holding. Tracks where things physically live (box / drawer), condition, and store stock vs. personal collection. Pushes updates directly to BrickLink via their API — listings and inventory sync live, no more manual XML uploads or fighting the BrickLink UI. Basically: the fast, dumb-simple flow I always wished BrickLink had — point, scan, price, sync.

The stack Next.js 14 + TypeScript + Tailwind for the app Zustand for state Rebrickable API for lookup, BrickLink API (OAuth 1.0a) for live two-way sync and pricing Camera-based visual part recognition built into the mobile shell Capacitor to wrap the web app into real iOS + Android apps from one codebase Firebase (auth, data, push) and Vercel for hosting StoreKit / Stripe for payments The part nobody warns you about: launching Writing the app was the easy 80%. Shipping it was the other 80%.

I started native in SwiftUI, then pivoted to web + Capacitor so one codebase serves web, iOS, and Android. Right call — but a real detour. Native plugins (StoreKit, Firebase messaging, the camera) didn't "just work" through the wrapper. I ended up writing patch scripts that run on every build to keep the iOS side stitched together. Then Apple's world: the Paid Apps agreement, banking and tax forms, server-side receipt validation, demo accounts for review, and the surprise that your first in-app purchase has to ship bundled with an app version. Plus two payment systems — Stripe on web, StoreKit on iOS, because Apple. The real lesson I can now build for my own problems.

That sentence sounds small, but sit with it. I had a niche, personal, "no one would ever build this for me" problem — and I solved it with software I made myself. The economics of that have completely flipped. It used to make no sense to build an app for an audience of one. With AI, it does.

So I keep coming back to this question: is the future of development thousands of small, bespoke apps — each one solving one person's specific problem? Not VC-backed startups chasing a billion users, but software the way you'd cook your own dinner. Personal. Disposable. Exactly fit to you.

AI didn't ship this app for me — Apple's process, the native glitches, the boring infrastructure still needed a human to grind through. But it lowered the barrier enough that "I wish this existed" can now become "so I built it" in a weekend. That changes who gets to make software. Maybe it changes everything.

It's also the lens I bring to Valdris work: the gap between "an AI can draft this" and "this is actually live and earning its keep" is where most of the real value — and most of the effort — still lives.

Try it iOS — Download on the App Store Web — brickstack.app.valdris.net Android — coming soon If you're a LEGO reseller or collector buried in bricks, I'd love your feedback. And if you're building bespoke apps for your own problems — tell me what you've made.

LEGO is a trademark of the LEGO Group. BrickStack and Valdris are not sponsored, authorized, or endorsed by the LEGO Group.