A small thesis
Print-on-demand has been stuck in the same loop for years: pick a template, swap a name, ship a t-shirt. It works, and it stops short of the thing that made bespoke goods worth paying for in the first place: the sense that a product was made for the person holding it.
Customei starts somewhere else. Rather than bolting a configurator onto a product page, the personalization canvas is a theme app block inside the product template, and it renders as the shopper types.
What that changes
- The canvas renders the real print area, so the geometry a shopper approves is the geometry the printer receives.
- The same rendering code runs in three places: the storefront runtime, the template editor, and print export. A layer cannot look one way in the editor and another on the press.
- The merchant owns the template. The shopper never sees a design tool, only their own product.
Ten layers that are not text-on-a-mug
Generative layers are the reason the canvas has to be live rather than a still preview: they compute their own artwork from what the shopper enters.
QR codes, calendars, maps, star maps, moon phases, string art, crosswords, word searches, word clouds and Spotify codes. Ten layer types, from the Studio plan up.
A crossword is the clearest case. The shopper types a word list, and the layout has to be solved before anything can be drawn. The solver the storefront uses is the same class the editor and print export use, and it settles a 28-word grid in around 1.3ms, fast enough that there is no reason to make anyone wait on a server round trip to see their own grid.
The rest is scaffolding
Pricing, fulfilment, proofing, routing: all of it exists so that canvas survives contact with production. That is the harder half of the problem, and it is why the print pipeline gets as much attention around here as the editor does.
