ArtPrompts Generator is a free, independent resource for anyone who writes prompts for AI image models. We build and maintain a browser-based prompt generator, and we publish plain-spoken guides that explain how to get better results from Midjourney, Stable Diffusion, DALL·E, Flux and Niji. There is no course to buy and no login to create — the tool and the guides are open to everyone.

Why we exist

Prompting looks simple until you try to get a specific result. The gap between “a cat, oil painting” and an image that actually matches the picture in your head is filled with vocabulary, model quirks and settings that almost nobody explains clearly. Most of what circulates online is either a wall of magic keywords with no reasoning, or a screenshot with no reproducible prompt behind it. We started ArtPrompts Generator to close that gap: to show the why behind a technique, not just hand over a string of words.

What we do

Two things, mainly.

  • The generator. Our on-site tool helps you assemble a structured prompt from a subject, a style, lighting, composition and model-specific parameters. It is a scaffold for thinking, not a black box — you can edit every part and copy the result straight into your model of choice.
  • The guides. We write reference material and step-by-step walkthroughs: what stylize and CFG actually change, how to build a negative prompt that helps rather than hurts, how to hold a consistent character across a series, and how the same idea has to be phrased differently for each model.

Who it’s for

If you are a hobbyist making art for the joy of it, a designer roughing out concepts, a marketer producing visuals on a deadline, or a curious beginner who typed one prompt and wondered why the result was nothing like you imagined — this site is for you. We assume no prior knowledge and we try never to hide behind jargon we haven’t defined.

The team

We are a small, independent editorial team, not a venture-backed company with a marketing department. Elena Vasquez, our founder and creative director, spent a decade art-directing campaigns before turning that eye toward generative models; she sets our editorial standard. Kai Nakamura, our lead prompt engineer and a former ML engineer, handles the technical side — samplers, CFG, token weighting and the settings that quietly change everything. Noah Bennett, our staff writer, produces the guides and prompt libraries, drawing on tens of thousands of images he has generated across Midjourney and Niji. You can read fuller bios on our Meet the Team page.

Our promise

We test techniques before we recommend them — you can read exactly how on our How We Test page — and we keep our guidance honest about the fact that AI models change constantly. A prompt that sang last month can fall flat after a version bump, and we say so when it happens. If you spot something that no longer works, or simply want to talk shop, we would love to hear from you via our contact page.