AI image prompt builder with token-weighting strategy

Build any image by assembling these blocks. Order matters — here is why:

[SUBJECT + KEY DETAIL] + [ACTION/POSE] + [ENVIRONMENT] + [ART STYLE] + [LIGHTING] + [COMPOSITION] + [COLOR PALETTE] + [DETAIL/QUALITY TAGS] + [ASPECT RATIO]

How token weighting works: Most generators give the strongest influence to the FIRST tokens. Your subject should always lead. Style and lighting in the middle shape the feel. Quality tags at the end act as a global polish layer. If you bury your subject after a long environment description, the scene will overpower the subject.

Example 1 — Painterly: 'A red fox curled asleep on a moss-covered stone, ancient pine forest, soft morning fog, storybook gouache illustration, warm side light, centered composition, amber and forest-green palette, intricate brushwork detail --ar 16:9'

Example 2 — Cinematic Photo: 'A street musician playing saxophone, rain-soaked Tokyo alley at night, neon reflections on wet pavement, photorealistic, backlit by a single warm streetlamp, shallow depth of field, teal and amber palette, 8k sharp detail --ar 21:9'

When results miss your intent:
• Subject wrong? Move it to the very first words and simplify adjectives.
• Style bleeding? Separate style keywords with commas — 'oil painting, not watercolor' or use (parentheses for emphasis) in Stable Diffusion.
• Too cluttered? Cut environment detail. Generators fill empty space creatively when you leave it sparse.
• Colors off? Name specific colors ('cerulean blue') instead of moods ('cool tones'). Explicit beats implicit.
• Composition ignored? Add 'rule of thirds' or 'centered subject' AND match it to your aspect ratio.