Dreams

What we are trying to make true. Not a roadmap — there are no dates here, and nothing on this page has happened yet.

We are not trying to make another inspiration library. Those exist and they are good at what they do. The gap is that the knowledge inside a visual language — what it requires, what it refuses, and how confident anyone should be about either — is almost never written down in a form a person or a machine can act on.

The ambitions

  1. Become a shared standard for visual language

    A vocabulary designers, developers, clients and machines can use to communicate visual intent without the word “editorial” meaning four incompatible things to four people in the same meeting.

  2. Become part of design education

    Useful enough to be referenced in art schools, design programmes and classrooms — not because we declare ourselves authoritative, but because the underlying knowledge earns it. That means citations, stated uncertainty and a record of what we got wrong.

  3. Become infrastructure

    Aesthetary knowledge available inside the creative and development tools people already use, rather than as another site to remember to visit.

  4. Give developers access to visual intelligence

    Structured visual-language knowledge through APIs, agent tools and machine-readable specifications, so software can reason about visual language instead of imitating screenshots of it. Every record already publishes as JSON; that is the first step, not the destination.

  5. Build a real creative community

    Designers contributing knowledge, creating recipes, remixing one another’s work, and receiving attribution when their ideas travel. Attribution designed in from the first version, because attribution retrofitted is attribution already lost.

  6. Reach people who need it

    We will publish concrete adoption goals when we have targets worth publicly owning. Naming a number before then would be theatre.

What is actually true today

A reference layer that documents visual languages with their evidence attached, a partly working mixer that shows what the corpus can and cannot resolve about a combination, and machine-readable records for every entry. Everything above is beyond that line.

Explore the corpus · How it works · About

Feedback