You've learned how to craft prompts. Now let's look under the hood! In this lesson, you'll discover how a few key settings can make your images sharper, more accurate, or more creative.

Let's start with a simple prompt and build from there.

Now, let's enhance our prompt by adding a stylistic element, as you did previously. This will make it more specific.

Style is just one lever you can pull. The model you choose — and the reference images you bring in — can shift your results just as much. Let's look at both.

So far, you've been working with OpenArt SDXL — the default model. But OpenArt gives you access to other Stable Diffusion versions, each with its own strengths. Choosing the right model is one of the quickest ways to steer your results before you even write a word.

You now know how to pick the model that fits your vision. But what if the vision itself is hard to put into words? That's a different kind of stuck. OpenArt has a feature for it: Prompt from Image in the prompt box.

Instead of writing a prompt from scratch, you upload a reference image and Stable Diffusion generates a text prompt that describes it — which you can then use or edit to create something similar.

This is how to use it:

This is your prompt: "Textured image of a man in a red vest and glasses writing at a desk. Walls are completely covered in overlapping papers and notes. Warm, focused light from a blue desk lamp illuminates his workspace. Expressive linework, cross-hatching, post-impressionist style, earthy tones, moody atmosphere." It's also a great starting point for experimentation. Try swapping out individual elements — the subject, the lighting, the mood — and see how the image shifts.

Swapping elements in a prompt gives you creative control. But there's another kind of control that works in the opposite direction — not adding what you want, but removing what you don't.

Negative prompts help you exclude unwanted elements, styles, or colors from your image. Some of the most commonly used negative prompt terms in Stable Diffusion are:

Let's see the difference negative prompts make in practice. First, without any restrictions.

See the difference? The second image avoids the neon aesthetic entirely. Negative prompts give you more control by filtering out elements that don't fit your vision. Now let's put model choice and negative prompts together.

Great job! In this lesson, you covered a lot of ground: building and enhancing prompts, choosing the right model for the job, reversing a prompt from an image, and using negative prompts to filter out what doesn't belong. Next up — samplers. You'll learn how they influence how Stable Diffusion processes your prompt and how to use them to get even more consistent, intentional results.