AtomWords

The Visual Dictionary of AI Art Keywords

Hi, I'm musefeng, founder of AtomWords.
As a multidisciplinary designer working across web, UI, and graphic design, I spend most of my days making images with AI — experimenting, breaking things, occasionally landing on something that makes me stop and stare.

I've always believed creativity isn't one big idea that suddenly appears. It's more like a scattered collection of isolated dots — images you've seen, words you've come across, references you saved on a whim — sitting loosely in your head with no obvious connection. Until one day, quietly, a few of them link up. And there it is. That's what inspiration feels like.
So I've always pushed myself to take in more. To widen the view.

One image I keep coming back to — the hollow onion
There's an AI image I made that I keep coming back to. A hollow onion. It's my favorite thing I've ever generated. Not because it's technically impressive — but because it looks exactly like life. The destination is the same for everyone. The digging, the exploring along the way — that's the meaning.

I think about that image a lot when I work on AtomWords.

Why I built this
When you create with AI, you write prompts — sometimes long ones. What used to frustrate me: I could see the final image, I could read the full prompt, and I still couldn't tell which word was doing the visual work.

Was it cinematic? Liminal? Brutalist?

Hard to say. That blind spot bothered me.

So I built AtomWords — a visual dictionary for AI prompts. The logic is simple: one word, one image. I call these atom words — the words that carry the most visual weight in a prompt. See exactly what that word looks like when it meets the model. No guessing. No digging through endless threads. Just: this word → this result.

How AtomWords is different
Most prompt sites show you a wall of full prompts next to a wall of images — subjects, styles, parameters, and render settings all tangled into one giant blob. The visual signal is buried in the noise.

AtomWords takes the other approach. Each entry isolates one visual variable, so you can finally see what a single word does on its own — separate from everything else competing for the model's attention.

What it's for
AtomWords is a place to look things up — and a place to wander. Maybe you're searching for a specific style keyword. Maybe you're scrolling and you stumble across a word you've never used before and think — I need that.

Call it an AI visual dictionary. Call it a visual prompt wiki. Call it an AIGC keyword library. You get to define what it is.

Use it for inspiration, use it for research, or browse with no agenda. Slowly, those little creative dots start to multiply.

Hope it sparks something.
— musefeng, Founder of AtomWords