Image Prompt Guide
A provider-agnostic guide for crafting effective AI image generation prompts. Covers illustrations, covers, infographics, and general imagery.
CRITICAL: grep references/ for detailed style/layout/palette specs and prompt templates.
Core Principles
Good image prompts follow structure first, aesthetics second:
- Layout/Composition — Describe spatial arrangement, zones, flow direction
- Specific Content — Use actual data, labels, terms (not vague descriptions)
- Visual Relationships — How elements connect and relate
- Semantic Colors — Meaning-based color choices (red=warning, green=success)
- Style Characteristics — Line treatment, texture, mood
- Aspect Ratio + Quality — End with dimensions and detail level
Universal Prompt Suffix
Always append to prompts:
Clean composition with generous white space. Simple or no background. Main elements centered or positioned by content needs.
What to Avoid
- Vague descriptions ("a nice image", "something beautiful")
- Literal metaphor illustrations (visualize the concept, not the metaphor)
- Missing concrete labels/annotations
- Photorealistic human faces (use simplified silhouettes)
- Overly complex backgrounds
Two-Dimensional Design System
All image generation uses two independent dimensions that combine freely:
| Dimension |
Controls |
Question |
| Type/Layout |
Information structure |
How is content organized? |
| Style |
Visual aesthetics |
What does it look like? |
Example: infographic (type) × blueprint (style) = technical data visualization with engineering aesthetics.
Quick Reference: Type Selection
| Content |
Recommended Type |
| Data, metrics, comparisons |
infographic |
| Stories, emotions, atmosphere |
scene |
| Processes, workflows, how-to |
flowchart |
| A vs B, pros/cons |
comparison |
| Architecture, models, concepts |
framework |
| History, evolution, milestones |
timeline |
| Product launch, announcements |
hero (cover) |
| Multiple topics overview |
bento-grid (infographic) |
| Abstract concept visualization |
conceptual (cover) |
Quick Reference: Style Selection
| Content Tone |
Recommended Style |
| Knowledge, tutorials, tech |
vector-illustration or blueprint |
| SaaS, productivity |
notion (minimal line art) |
| Personal, lifestyle |
warm or watercolor |
| Opinion, editorial, cultural |
screen-print or editorial |
| Education, teaching |
chalkboard or craft-handmade |
| Business, professional |
elegant |
| Fun, casual, playful |
flat-doodle or kawaii |
| Retro, nostalgic |
pixel-art or retro |
Quick Reference: Color Palette
| Mood |
Palette |
Key Colors |
| Friendly, warm |
warm |
Orange #ED8936, Golden #F6AD55, Cream bg |
| Professional |
elegant |
Slate #475569, Gold #D4AF37, Ivory bg |
| Technical |
cool |
Blue #2563EB, Teal #0D9488, Light bg |
| Premium, cinematic |
dark |
Charcoal #1A1A2E, Purple #7C3AED, Dark bg |
| Natural, organic |
earth |
Green #059669, Brown #92400E, Warm bg |
| Bold, energetic |
vivid |
Red #EF4444, Cyan #06B6D4, White bg |
| Gentle, creative |
pastel |
Pink #FCA5A5, Lavender #C4B5FD, Soft bg |
| Zen, minimal |
mono |
Grays + one accent, White bg |
Scenario Templates
Infographic / Data Visualization
[Title] - Data Visualization
Layout: [grid/radial/hierarchical]
ZONES:
- Zone 1: [data point with specific values]
- Zone 2: [comparison with metrics]
- Zone 3: [summary/conclusion]
LABELS: [specific numbers, percentages, terms]
COLORS: [semantic color mapping]
STYLE: [style characteristics]
ASPECT: 16:9
Scene / Atmospheric
[Title] - Atmospheric Scene
FOCAL POINT: [main subject]
ATMOSPHERE: [lighting, mood, environment]
MOOD: [emotion to convey]
COLOR TEMPERATURE: [warm/cool/neutral]
STYLE: [style characteristics]
ASPECT: 16:9
Flowchart / Process
[Title] - Process Flow
Layout: [left-right/top-down/circular]
STEPS:
1. [Step] - [description]
2. [Step] - [description]
CONNECTIONS: [arrow types, decision points]
STYLE: [style characteristics]
ASPECT: 16:9
Comparison
[Title] - Comparison View
LEFT SIDE - [Option A]:
- [Point 1]
- [Point 2]
RIGHT SIDE - [Option B]:
- [Point 1]
- [Point 2]
DIVIDER: [visual separator]
STYLE: [style characteristics]
ASPECT: 16:9
Cover Image
[Title] - Cover Image
TYPE: [hero/conceptual/typography/metaphor/scene/minimal]
FOCAL ELEMENT: [main visual]
TITLE TEXT: [exact title, language]
TEXT LEVEL: [none/title-only/title-subtitle]
PALETTE: [color scheme]
RENDERING: [flat-vector/hand-drawn/painterly/digital]
MOOD: [subtle/balanced/bold]
ASPECT: 16:9
Composition Principles
| Principle |
Guideline |
| White space |
40-60% breathing room |
| Visual anchor |
Main element centered or offset left |
| Characters |
Simplified silhouettes, NOT realistic humans |
| Text |
Large, prominent, handwritten fonts preferred |
| Title |
Use exact title from source, never invent |
| Hierarchy |
Size = importance |
References
references/styles.md — 28+ visual styles with descriptions, color palettes, and compatibility matrixreferences/layouts.md — 21 infographic layouts with structure, best-for, and visual elementsreferences/palettes.md — 10 color palettes with hex values, decorative hints, and use casesreferences/templates.md — Detailed prompt templates with style-specific overridesreferences/auto-selection.md — Content signal → dimension mapping rulesreferences/sd-styles.csv — 1062 image generation style presets (name, prompt, negative_prompt) merged from 7 sources (Douleb 850+, Fooocus 200+, BBrother, rocketingdatascience, twri/sdxl_prompt_styler, jheidt)references/nano-banana.md — Nano Banana (Gemini 2.5 Flash) prompt guide: core principles, 7 style templates (3D chibi, product ads, retro, effects, style transfer, design, creative scenes), structured JSON prompts, mixed-style techniquesreferences/nano-banana-cases.csv — 100 real-world Nano Banana/GPT-4o prompt cases with zh/en bilingual prompts (from awesome-nano-banana)
1---2name: image-prompt-guide3description: USE THIS SKILL WHEN writing prompts for AI image generation: creating illustrations, cover images, infographics, diagrams, or any visual content. Covers prompt structure, visual styles, color palettes, layout types, composition principles, scenario-specific templates, SD style presets, and Nano Banana (Gemini Flash) prompt techniques with 100 real-world cases. Triggers on: image prompt, generate image, illustration, cover image, infographic, 配图, 封面图, 信息图, dall-e prompt, midjourney prompt, nano banana, gemini image, image generation.4---56# Image Prompt Guide78A provider-agnostic guide for crafting effective AI image generation prompts. Covers illustrations, covers, infographics, and general imagery.910CRITICAL: grep `references/` for detailed style/layout/palette specs and prompt templates.1112## Core Principles1314Good image prompts follow **structure first, aesthetics second**:15161. **Layout/Composition** — Describe spatial arrangement, zones, flow direction172. **Specific Content** — Use actual data, labels, terms (not vague descriptions)183. **Visual Relationships** — How elements connect and relate194. **Semantic Colors** — Meaning-based color choices (red=warning, green=success)205. **Style Characteristics** — Line treatment, texture, mood216. **Aspect Ratio + Quality** — End with dimensions and detail level2223### Universal Prompt Suffix2425Always append to prompts:26> Clean composition with generous white space. Simple or no background. Main elements centered or positioned by content needs.2728### What to Avoid29- Vague descriptions ("a nice image", "something beautiful")30- Literal metaphor illustrations (visualize the *concept*, not the metaphor)31- Missing concrete labels/annotations32- Photorealistic human faces (use simplified silhouettes)33- Overly complex backgrounds3435## Two-Dimensional Design System3637All image generation uses **two independent dimensions** that combine freely:3839| Dimension | Controls | Question |40|-----------|----------|----------|41| **Type/Layout** | Information structure | *How is content organized?* |42| **Style** | Visual aesthetics | *What does it look like?* |4344Example: `infographic` (type) × `blueprint` (style) = technical data visualization with engineering aesthetics.4546## Quick Reference: Type Selection4748| Content | Recommended Type |49|---------|-----------------|50| Data, metrics, comparisons | `infographic` |51| Stories, emotions, atmosphere | `scene` |52| Processes, workflows, how-to | `flowchart` |53| A vs B, pros/cons | `comparison` |54| Architecture, models, concepts | `framework` |55| History, evolution, milestones | `timeline` |56| Product launch, announcements | `hero` (cover) |57| Multiple topics overview | `bento-grid` (infographic) |58| Abstract concept visualization | `conceptual` (cover) |5960## Quick Reference: Style Selection6162| Content Tone | Recommended Style |63|-------------|------------------|64| Knowledge, tutorials, tech | `vector-illustration` or `blueprint` |65| SaaS, productivity | `notion` (minimal line art) |66| Personal, lifestyle | `warm` or `watercolor` |67| Opinion, editorial, cultural | `screen-print` or `editorial` |68| Education, teaching | `chalkboard` or `craft-handmade` |69| Business, professional | `elegant` |70| Fun, casual, playful | `flat-doodle` or `kawaii` |71| Retro, nostalgic | `pixel-art` or `retro` |7273## Quick Reference: Color Palette7475| Mood | Palette | Key Colors |76|------|---------|------------|77| Friendly, warm | `warm` | Orange #ED8936, Golden #F6AD55, Cream bg |78| Professional | `elegant` | Slate #475569, Gold #D4AF37, Ivory bg |79| Technical | `cool` | Blue #2563EB, Teal #0D9488, Light bg |80| Premium, cinematic | `dark` | Charcoal #1A1A2E, Purple #7C3AED, Dark bg |81| Natural, organic | `earth` | Green #059669, Brown #92400E, Warm bg |82| Bold, energetic | `vivid` | Red #EF4444, Cyan #06B6D4, White bg |83| Gentle, creative | `pastel` | Pink #FCA5A5, Lavender #C4B5FD, Soft bg |84| Zen, minimal | `mono` | Grays + one accent, White bg |8586## Scenario Templates8788### Infographic / Data Visualization8990```91[Title] - Data Visualization9293Layout: [grid/radial/hierarchical]9495ZONES:96- Zone 1: [data point with specific values]97- Zone 2: [comparison with metrics]98- Zone 3: [summary/conclusion]99100LABELS: [specific numbers, percentages, terms]101COLORS: [semantic color mapping]102STYLE: [style characteristics]103ASPECT: 16:9104```105106### Scene / Atmospheric107108```109[Title] - Atmospheric Scene110111FOCAL POINT: [main subject]112ATMOSPHERE: [lighting, mood, environment]113MOOD: [emotion to convey]114COLOR TEMPERATURE: [warm/cool/neutral]115STYLE: [style characteristics]116ASPECT: 16:9117```118119### Flowchart / Process120121```122[Title] - Process Flow123124Layout: [left-right/top-down/circular]125126STEPS:1271. [Step] - [description]1282. [Step] - [description]129130CONNECTIONS: [arrow types, decision points]131STYLE: [style characteristics]132ASPECT: 16:9133```134135### Comparison136137```138[Title] - Comparison View139140LEFT SIDE - [Option A]:141- [Point 1]142- [Point 2]143144RIGHT SIDE - [Option B]:145- [Point 1]146- [Point 2]147148DIVIDER: [visual separator]149STYLE: [style characteristics]150ASPECT: 16:9151```152153### Cover Image154155```156[Title] - Cover Image157158TYPE: [hero/conceptual/typography/metaphor/scene/minimal]159FOCAL ELEMENT: [main visual]160TITLE TEXT: [exact title, language]161TEXT LEVEL: [none/title-only/title-subtitle]162PALETTE: [color scheme]163RENDERING: [flat-vector/hand-drawn/painterly/digital]164MOOD: [subtle/balanced/bold]165ASPECT: 16:9166```167168## Composition Principles169170| Principle | Guideline |171|-----------|-----------|172| **White space** | 40-60% breathing room |173| **Visual anchor** | Main element centered or offset left |174| **Characters** | Simplified silhouettes, NOT realistic humans |175| **Text** | Large, prominent, handwritten fonts preferred |176| **Title** | Use exact title from source, never invent |177| **Hierarchy** | Size = importance |178179## References180181- `references/styles.md` — 28+ visual styles with descriptions, color palettes, and compatibility matrix182- `references/layouts.md` — 21 infographic layouts with structure, best-for, and visual elements183- `references/palettes.md` — 10 color palettes with hex values, decorative hints, and use cases184- `references/templates.md` — Detailed prompt templates with style-specific overrides185- `references/auto-selection.md` — Content signal → dimension mapping rules186- `references/sd-styles.csv` — 1062 image generation style presets (name, prompt, negative_prompt) merged from 7 sources (Douleb 850+, Fooocus 200+, BBrother, rocketingdatascience, twri/sdxl_prompt_styler, jheidt)187- `references/nano-banana.md` — Nano Banana (Gemini 2.5 Flash) prompt guide: core principles, 7 style templates (3D chibi, product ads, retro, effects, style transfer, design, creative scenes), structured JSON prompts, mixed-style techniques188- `references/nano-banana-cases.csv` — 100 real-world Nano Banana/GPT-4o prompt cases with zh/en bilingual prompts (from awesome-nano-banana)
wenerme