Texture Grid: Why I Built a Tool With No Tutorial, No Right Answer, No Progress Bar

Back in January, I kept circling one observation. Almost everything we do online runs through words. We read them, write them, react to them, argue about them in comment sections that exist purely to generate more words. Even the tools marketed as creative are word-first underneath. You type a prompt to generate an image. You caption a photo to make it findable. You write a review to prove you felt something. Language sits between you and almost every digital experience, doing the job of translating a feeling into something that can be typed, searched, and stored.

I wanted to try building something that skipped that translation. Not a tool where you describe how a word makes you feel, but one where you show it, through something that isn't language at all. Texture felt like the right material for that. It's specific, it's physical even when it's on a screen, and it doesn't come pre-loaded with the vocabulary everyone already agrees on for describing an emotion. Two people can look at the same word and reach for completely different textures, and neither of them is wrong, because there's no dictionary sitting between the word and the response.

That idea sat mostly unbuilt until International Women's Day, when Lovable offered a free twenty-four hour build window. I used it. Not because the timing was perfect, but because a genuinely free day to build something is rare enough that turning it down felt like the wrong instinct. Twenty-four hours is not long enough to build something finished. It is long enough to build something true to the idea, which turned out to matter more.

What came out of that day was Texture Grid: you're given a word, and a grid of textures, and asked to choose which ones feel like the word to you. No explanation of what a "correct" answer would look like, because there isn't one.

One of the first words I tested it with was DANGER. I picked four textures for it: peeling paint, brick, smoke emerging from rock face, and chain-link fence. It turns the person looking at it from a viewer into a participant before they've clicked anything. You look at four answers that aren't yours, and the question shifts from "what did she mean by this" to "what would I put there." That shift, from interpreting someone else's choice to making your own, is the whole point of the tool compressed into one image.

When I showed it to friends and family, the reaction told me more than I expected. A number of them were visibly uncomfortable. Not with the concept, they understood what it was asking, but with the experience of actually doing it. There was no tutorial, so nobody could hand them a set of instructions and let them follow along instead of deciding for themselves. There was no right answer, so nobody could aim for a correct response and feel the small relief of having got it right.

There was no progress bar, so nobody could measure how much was left or feel the momentum of moving toward a finish line. All three of those things are so standard in digital products that their absence registers as a kind of foreignness, even though none of them are actually necessary to the task. They're scaffolding for passive use, and once you take them away, the only thing left to do is respond.

That discomfort is the most useful data I've gotten from the whole project so far. Most online experiences are engineered to reduce friction, to make sure nobody ever has to sit with an open-ended question and generate their own answer from nothing. Texture Grid does the opposite on purpose. The friction isn't a bug I haven't fixed yet. It's the entire mechanism. If choosing a texture for an abstract word felt effortless, it would mean the tool had quietly reintroduced a right answer somewhere, a most-popular choice, a suggested response, something to lean on instead of actually deciding. The discomfort is what participation feels like when there's nothing else to fall back on.

I keep thinking about what this could mean for institutions that already ask people to feel something they can't easily describe. A museum visitor standing in front of an object doesn't usually walk away with words for what just happened. Neither did I, most of the times a piece of art has actually stopped me. What would it look like for that visitor to answer through texture instead of through a comment box or a star rating, the same way the friends I tested this on had to answer without a script to follow. I don't have that built yet. But it's the same question in a different room.

For now, Texture Grid exists as one page, unpolished, open to anyone who wants to try it. I'd rather have people from anywhere in the world test it in its current state than wait until it looks finished, because the thing I'm actually curious about isn't whether it's polished. It's what happens to people when participation is the only option on the table.

Try it here: https://texturegr.id/