First Steps Into BoundingBox.store

I’ve wanted computer-vision merch for years: something that actually spoke the language of the field, bounding boxes, confidence scores, detection overlays, instead of generic “AI” clipart on a t-shirt. I looked, on and off, for a long time. I never found it. So BoundingBox.store is me finally just building it.

Why start here

Everything else on this site came out of research or a job: a thesis, a paper, a role. This one didn’t. It’s the first project I’ve started purely because I wanted the thing to exist, not because a lab or a client needed it. That changes the shape of the work: no advisor, no client scope, no deadline but the one I set myself, just an idea I kept coming back to until building it felt easier than continuing to want it.

The idea itself is deliberately narrow: one concept, the bounding box, drawn “with the precision of a trim mark, not the noise of a heads-up display,” running through everything, the copy, the product photography, even the loading state. A tee, a hoodie, a sticker pack, and an API-shaped joke of a product, all wrapped in the same visual language I spent a Ph.D. staring at.

Building it

This is also my first real e-commerce build, and my first time reaching for Next.js instead of Astro on a personal project, mostly because I wanted a proper headless setup: Shopify lined up as the catalog and checkout backend, a custom storefront handling everything about how it looks and feels. That part is still being built behind the scenes, so what’s actually live right now is the first piece of it: a pre-launch landing page carrying the same bounding-box visual language, with a waitlist so people can hear about the first drop before it exists.

Tech stack

Where it stands

It’s pre-launch: the landing page and waitlist are live at boundingbox.store, the catalog and checkout are still in development. I expect the design to keep evolving once I’m actually using the thing rather than just planning it. I wrote up the build itself, the design system, the tech choices, in more detail on the project page. This post is more about the “why now” than the “how”: sometimes the fastest way to get the thing you’ve been looking for is to stop looking and build it.