json image prompting

Using JSON to structure AI image prompts can enhance clarity and control over the generated visuals. This method is particularly effective with models like GPT-4o, DALL·E, and Midjourney.(glenegrant.com, Medium)


🧠 Benefits of JSON-Structured Prompts

  • Precision: Clearly defines each element of the image.

  • Consistency: Maintains uniformity across multiple prompts.

  • Automation: Facilitates programmatic generation and manipulation of prompts.(DocsBot AI)


🧰 Common JSON Fields for Image Prompts

Field Description Example
scene Overall setting or environment "a bustling futuristic city at night"
subjects Key elements or characters in the image [{"type": "robot", "description": "sleek design with neon lights"}]
style Artistic style or genre "cyberpunk"
color_palette Dominant colors used in the image ["neon blue", "electric purple", "black"]
lighting Lighting conditions "glowing neon signs and ambient streetlights"
mood Emotional tone of the image "energetic and mysterious"
background Details about the background setting "skyscrapers with animated billboards"
composition Arrangement of elements within the frame "rule of thirds"
camera Camera angle and focus details {"angle": "low", "focus": "sharp on subject"}
medium Medium or technique used "digital illustration"
textures Surface qualities of elements ["metallic", "glossy"]
resolution Desired resolution or output size "4K"
details Additional specific attributes {"weather": "light rain", "accessories": "holographic umbrella"}
effects Special visual effects ["motion blur", "lens flare"]
inspirations References to artists or styles ["inspired by Blade Runner aesthetics"]

🎨 Example JSON Prompt

{
  "scene": "a bustling futuristic city at night",
  "subjects": [
    {
      "type": "robot",
      "description": "sleek design with neon lights",
      "pose": "walking through the crowd",
      "position": "foreground"
    }
  ],
  "style": "cyberpunk",
  "color_palette": ["neon blue", "electric purple", "black"],
  "lighting": "glowing neon signs and ambient streetlights",
  "mood": "energetic and mysterious",
  "background": "skyscrapers with animated billboards",
  "composition": "rule of thirds",
  "camera": {
    "angle": "low",
    "focus": "sharp on subject"
  },
  "medium": "digital illustration",
  "textures": ["metallic", "glossy"],
  "resolution": "4K",
  "details": {
    "weather": "light rain",
    "accessories": "holographic umbrella"
  },
  "effects": ["motion blur", "lens flare"],
  "inspirations": ["inspired by Blade Runner aesthetics"]
}


🔗 Additional Resources

  • Prompt Gallery by Google Cloud: Explore various prompt examples for generative AI models. (Google Cloud)

  • DocsBot AI Prompt: Generate images or detailed descriptions based on JSON data. (DocsBot AI)(DocsBot AI)

  • Prompt Structuring with JSON: Learn how JSON structuring improves AI creative outputs. (glenegrant.com)(glenegrant.com)