You've already taken in most of these movements without the names attached. Swiss grids on signage, Memphis on a coffee cup, Bauhaus geometry on half the posters in a museum shop.
What's missing usually isn't the look. It's knowing which is which, and what any of them were arguing about. So this works from the other end. Make something first, then find out what you made. You put a sheet together and it gets read against fifteen movements while you work.
It isn't here to tell you what your work is. It puts a name to whatever instinct you were already following, gives you somewhere to start reading, and writes you a prompt you can hand straight to an AI model so the thing you stumbled into can go on being useful.
Make → Discover → Learn → Make again
There are eight layers: the grid, a mass, the ink, the process, the surface, a headline, ornament, and the tilt. You can switch any of them off, add another mass, or go back and change the grid long after you placed the headline. Nothing is waiting on anything else, and there's no order you're supposed to work in.
Drag on the sheet itself or use the controls beside it. Both do the same job. The reading keeps up as you go, so you never have to call yourself finished to see it.
Every choice shifts four numbers: how ordered the sheet is, how ornamented, how warm, how loud. Each of the fifteen movements sits somewhere in that same space. The reading is whichever one you're closest to, plus the runner up, because the disagreement between two of them usually tells you more than the winner does.
It's a model, and models are rough. Two movements can sit almost on top of each other here and still be a century and a real argument apart in life. Take it as a place to start reading, not a result.
Every reading has a second tab, and it is the part you actually leave with. It's a written prompt, built from the sheet in front of you, ready to paste into Claude or ChatGPT or whatever you work with. Not a label. Something to build the next thing from.
It never asks for your poster back. It writes out a direction: the movement you landed on and the one running underneath it, the four inks with a job assigned to each and body text deliberately left off the list, how the composition sits, what the type is doing, and the process the sheet came out of. Then it says which of that won't survive the jump to a screen, and what to keep when the process fights the medium. Point it at a site, an identity, a deck, a piece of packaging.
The second version returns the same decisions as UI tokens instead: hex values, radius, spacing step, measure, type scale, how buttons and cards and focus rings should behave, and the contrast checks that go with them. Both copy in one click, and both rewrite themselves as you edit, so the prompt always describes the sheet as it currently stands rather than one you had a minute ago.
A movement isn't a colour scheme. What separates Dada from Swiss design isn't which inks they liked. It's how the thing got made.
So a sheet can come out as a photograph screened into dots, or paper cut up and pasted down with the torn edges showing, or the compass work left visible under the artwork, or colour plates pulled slightly out of register, or brush marks with the pressure still in them, or a photocopy that's been copied too many times. The headline can be set in type, or built letter by letter out of scraps.
That's what stops this being a colour filter, and it's why no two movements here should ever hand you the same sheet in different colours.
Every sheet in the gallery is a real result rather than a mock-up. Each one was made with the rules you get, then tuned until it genuinely reads as its own movement, and pushed away from whatever the others had already taken so you don't end up with two that look alike. Open any of them for where it came from, what it was arguing for, and which movements it grew out of or turned against.
They lean heavily European and North American. That's a fair criticism of this first pass. One Japanese tradition against fourteen Western ones isn't a map of how the world has made things look, and widening it is the next thing worth doing.
It's all arithmetic. No model, no dataset, nothing calling out to an API. Every poster is drawn in your browser from a seed, so none of it is stock imagery and no two sheets come out the same. The movement coordinates and the histories were researched and placed by hand.
A personal project, made out of curiosity about why things look the way they do, and where those ideas came from.