Image Prompts

Image Prompts

Generate optimized AI image prompts for popular image generation platforms including Midjourney, DALL-E, Stable Diffusion, and Leonardo AI.

Capabilities

• Multi-Platform Support: Generate prompts optimized for Midjourney, DALL-E, Stable Diffusion, and Leonardo AI
• Style-Aware Prompts: 15+ art styles with optimized keywords (photorealistic, anime, oil painting, cyberpunk, etc.)
• Subject Enhancement: Automatically enhance subject descriptions with relevant modifiers
• Parameter Suggestions: Platform-specific parameters (aspect ratio, quality, stylize, etc.)
• Prompt Variations: Generate multiple variations of prompts for A/B testing
• Negative Prompts: Auto-generate appropriate negative prompts per style
• Template System: Pre-built templates for common use cases
• History Tracking: SQLite storage for generated prompts

Installation

cd skills/image-prompts
npm install
npm run build

Usage

CLI

# Generate prompts for a subject
npx image-prompts generate "a mystical forest" --style fantasy --mood mysterious

# Generate with specific platform settings
npx image-prompts g "cyberpunk city" -p midjourney -s cyberpunk -m futuristic -l neon -a 16:9

# Use a template
npx image-prompts template character-portrait --var '{"subject":"a brave warrior"}'

# List available templates
npx image-prompts templates

# View prompt history
npx image-prompts history --limit 10

# List available options
npx image-prompts styles
npx image-prompts moods
npx image-prompts platforms

Library

import { ImagePromptsSkill } from '@openclaw/image-prompts';

const skill = new ImagePromptsSkill();

// Generate prompts
const prompts = await skill.generatePrompts({
  subject: 'a mystical forest with ancient trees',
  platform: 'midjourney',
  style: 'fantasy',
  mood: 'mysterious',
  lighting: 'volumetric',
  quality: 'ultra',
  aspectRatio: '16:9',
  count: 3
});

console.log(prompts[0].prompt);
// "fantasy art, magical, ethereal, mythical, enchanted, dreamlike of a mystical forest with ancient trees, mysterious, enigmatic, shadowy, intriguing, atmospheric, volumetric lighting, god rays, atmospheric lighting, ultra detailed, 8k, hyperrealistic, maximum quality, intricate details"

console.log(prompts[0].negativePrompt);
// "blurry, low quality, distorted, deformed, ugly, bad anatomy, watermark..."

console.log(prompts[0].parameters);
// { version: "6", stylize: 100, chaos: 0, ar: "16:9" }

// Get formatted command for Midjourney
const command = skill.formatPromptWithParameters(
  prompts[0].prompt,
  'midjourney',
  prompts[0].parameters
);
console.log(command);
// "fantasy art, magical... --ar 16:9 --stylize 100 --v 6"

// Use templates
const template = await skill.getTemplate('character-portrait');
const prompt = skill.applyTemplate(template, {
  subject: 'a brave warrior',
  style: 'digital art',
  mood: 'dramatic',
  lighting: 'cinematic',
  quality: 'ultra detailed',
  perspective: 'close up'
});

await skill.close();

Art Styles

• photorealistic: Ultra-realistic photography style
• digital-art: Modern digital illustration
• oil-painting: Classic oil painting aesthetic
• watercolor: Soft watercolor wash effects
• pencil-sketch: Hand-drawn pencil sketch
• anime: Japanese anime/manga style
• 3d-render: 3D computer generated imagery
• pixel-art: Retro pixel art style
• concept-art: Professional concept art
• cyberpunk: Futuristic cyberpunk aesthetic
• fantasy: High fantasy art style
• minimalist: Clean minimalist design
• surrealism: Surreal dreamlike imagery
• impressionist: Impressionist painting style
• pop-art: Bold pop art style

Moods

peaceful, dramatic, mysterious, joyful, melancholic, epic, whimsical, dark, romantic, futuristic

Lighting Options

natural, studio, dramatic, soft, golden-hour, neon, cinematic, backlit, volumetric, moonlight

Camera Perspectives

eye-level, birds-eye, worms-eye, close-up, wide-angle, macro, portrait, aerial, isometric, dutch-angle

Supported Platforms

Midjourney

Parameters: version, stylize, chaos, weird, tile, no

DALL-E

Parameters: quality (standard/hd), style (vivid/natural), size

Stable Diffusion

Parameters: steps, cfg_scale, seed, sampler, width, height

Leonardo

Parameters: photoReal, alchemy, guidance_scale

Generic

No platform-specific parameters

Templates

Built-in templates for common use cases:

• character-portrait: Detailed character portrait with personality
• landscape-scene: Breathtaking landscape environment
• product-shot: Professional product photography
• abstract-art: Abstract artistic composition
• architectural: Architectural visualization

Database

Prompts are stored in SQLite at ~/.openclaw/skills/image-prompts/prompts.db:

• generated_prompts: All generated prompts with metadata
• prompt_templates: Reusable prompt templates
• style_presets: Custom style configurations

No External Dependencies

This skill uses local prompt generation patterns and does not require external API keys. It works entirely offline.