Reference · How-to · ~8 min

How to create an image with AI

Good images come from **specific prompts** — subject, style, lighting, composition.

Steps

1. Pick a tool (DALL·E, Midjourney, Stable Diffusion, etc.)

2. Describe **subject + setting + style + mood**

3. Add negatives if supported: "no text, no watermark, no extra fingers"

4. Generate 3–4 variants; iterate on the best

5. Check **usage rights** for commercial work

[Subject], [action/pose], [setting],
[art style / medium], [lighting],
[color mood], [camera/framing]
Avoid: blurry, distorted hands, unreadable text

A cozy reading nook by a rainy window, watercolor illustration,
warm lamp light, soft blues and amber, wide shot, children's book style

Text in images is often garbled

Faces/hands may need retries

Don't use living artists' names without permission policies

**Try the lesson:** `creative-tools` in Lane B