Use Cases
Illustration in a consistent visual language for books, magazines and digital publishing.
Short answer
For AI illustration use Midjourney or Leonardo AI and describe the style by technique rather than by artist name: "heavy outlines, flat colour fields, limited palette, engraved texture". For consistency across a series, use Midjourney's style reference (--sref) parameter or Leonardo's saved style; repeating the prompt alone is not enough.
Illustration is one of the areas where AI tools are strongest — Midjourney's aesthetic interpretation is an advantage here rather than a flaw.
But the real difficulty in professional illustration is not producing one beautiful image. The thirty pages of a book, the six articles in a magazine or the twenty screens of an app need images in the SAME visual language. AI's natural tendency is the opposite.
This page focuses on the methods for moving from individual beauty to series consistency.
Instead of naming an artist, describe the technique: line weight, colour count, texture, shading method. That is both legally safer and more controllable in practice.
Save the first illustration you like as a reference and bind the rest of the series to it. For a book or magazine, consistency matters more than the quality of individual images.
Illustrations going to print need 300 DPI. For flat, outlined illustration Recraft's vector output is right; for textured, painterly work an upscaler is the way.
Run the same prompt twice and you get two different images. That is richness for a single image and a disaster for a series.
The most reliable fix is style reference. In Midjourney you pass the link of an image you like with the --sref parameter; new generations inherit its colour, texture and line character while the subject changes. Leonardo AI's saved styles do the same job.
The second layer is making the style portion of the prompt a fixed block. Write subject and style separately, and change only the subject on each generation. Naming the palette colours explicitly ("ochre, teal, cream") is that block's most effective part.
Series with characters carry an extra problem: keeping the same character looking the same across pages. Midjourney's character reference (--cref) parameter is built for this; on the Stable Diffusion side, training a LoRA for the character itself is more powerful but more laborious.
Naming a living artist in a prompt is the known shortcut to their style, and that is precisely why it is a problem.
The legal position is unsettled but the risk is real: using an artist's name in commercial generation is ethically indefensible and breaches some platforms' terms. Most tools have already begun filtering living artists' names.
The practical argument is more persuasive: an artist's name gives you that artist's average, not control. Describe the technique instead — line weight, colour count, texture, shading method — and you can tune the result. A description like "heavy black outlines, flat colour fields, three-colour palette, paper texture" is both safer and more useful than a name.
Save the first illustration you like as a style reference and bind the whole series to it: Midjourney's --sref parameter and Leonardo AI's saved styles both do this. Also make the style portion of your prompt a fixed text block and change only the subject. If a character recurs, use Midjourney's --cref parameter.
Do not use living artists' names. It is ethically indefensible, breaches some platforms' terms and most tools already filter those names. It is also weak in practice: a name gives you that artist's average, while describing the technique lets you tune the result.
Print needs 300 DPI. For flat, outlined illustration the best route is Recraft's vector (SVG) output — it stays sharp at any size. Vector is not suitable for textured, painterly work; generate those at the highest resolution available, run them through an upscaler and inspect at one hundred per cent zoom before printing.
Prompts are written in English because most tools give markedly better results with English prompts.
editorial illustration, a person climbing a staircase made of open books, limited palette of ochre teal and cream, heavy black outlines, flat colour fields with subtle paper grain, mid-century poster technique, generous negative space
Works well in these tools: midjourney, leonardo-ai, recraft
Note: Naming the palette colours explicitly is the simplest and most reliable way to hold a series together.
children's book illustration, a small fox looking up at a paper lantern in a forest clearing, soft watercolour washes, warm dusk palette, gentle rounded shapes, visible paper texture, no outlines, storybook composition
Works well in these tools: midjourney, leonardo-ai, recraft
Note: Character consistency is mandatory in children's books; save the first character you generate as a reference and use it on every page.
The subscription generator that delivers the most consistent artistic and aesthetic results.
Free credits that renew daily, a large model library and dedicated tooling for game assets.
Generates logos, icons and illustrations directly as scalable vectors.
The strongest open-weight family for photorealistic people and for text rendered inside images.