×
AI Image Generator · Prompt Engineering

The Secret JSON Prompt System That Can Make Your AI Images Sell

The problem when we generate image is that we use it like a slot machine, we keep prompting and prompting, hoping the AI spit out an image that’s close to our imagination. You need to engineer prompts, just like cinematographers plan a scene.

Thai An Le Thai An Le September 17 5 min read 28 0 0
The Secret JSON Prompt System That Can Make Your AI Images Sell

If you’ve ever typed something like:

“clay dog on moon”

…and got back something totally different from what you imagined, that looks “AI”, it’s not the AI’s fault, it’s your prompt.

The problem when we generate image is that we use it like a slot machine, we keep prompting and prompting, hoping the AI spit out an image that’s close to our imagination.

You need to engineer prompts, just like cinematographers plan a scene.

Why This Isn’t Just About Pretty Pictures

People are already using AI-generated images to make money in:

However, Most AI art isn’t consistent enough to sell. Buyers want a look that’s repeatable. That’s where structured JSON prompts come in.

Why This Works Like Magic

The Prompt Blueprint AI Whisperers Use

Copy this. Keep it. Share it. This is your new master key:

{
  "image_purpose": "still_life | portrait | landscape | editorial | concept_art",
  "scene": {
    "environment": "studio | interior | exterior | on_location",
    "background": {
      "material": "painted wall | seamless paper | fabric | natural scenery",
      "color_tone": "neutral-warm | neutral-cool | vivid | dark",
      "cleanliness": "clean | textured | aged"
    },
    "surface": {
      "material": "wood | stone | glass | metal",
      "tone": "light | medium | dark",
      "finish": "matte | satin | glossy"
    }
  },
  "composition": {
    "framing": "horizontal | vertical | square | panoramic",
    "tiers": 1,
    "arrangement_notes": "grid | staggered | pyramid | asymmetric",
    "negative_space": "tight | medium | airy",
    "camera_height": "eye-level | slightly above | slightly below"
  },
  "camera": {
    "focal_length_mm": 35,
    "aperture_f": 4,
    "lens_type": "prime | zoom",
    "focus_strategy": "all subjects sharp | selective",
    "distance_descriptor": "close | mid | wide"
  },
  "lighting": {
    "type": "natural | strobe | continuous",
    "direction": "left | right | back | top",
    "quality": "diffused | soft | hard",
    "contrast": "low | medium | high",
    "color_temperature": "neutral | warm | cool",
    "shadow_behavior": "soft edge | defined"
  },
  "color_grading": {
    "palette_keywords": ["warm beige", "soft teal"],
    "saturation_level": "muted | natural | vivid",
    "overall_warmth": "cool | neutral | warm"
  },
  "materials_and_texture": {
    "primary_materials": ["ceramic", "wood"],
    "surface_finish": "unglazed | glossy | satin",
    "texture_notes": "fine grit, brushed lines"
  },
  "mood": ["calm", "earthy", "inviting"],
  "imperfections": {
    "include": true,
    "types": ["uneven rims", "finger depressions", "tiny nicks"]
  },
  "subjects": [
    {
      "id": "s1",
      "category": "vase | bowl | figure | other",
      "tier": "upper | lower",
      "pose_or_orientation": "front-facing | 3/4 | profile",
      "silhouette": "cylindrical | rounded | irregular",
      "special_features": [],
      "proportion_notes": "",
      "size_relative": "small | medium | large",
      "condition": "handmade irregularities emphasized"
    }
  ],
  "depth_of_field": {
    "subject_sharpness": "all subjects crisp",
    "background_softness": "subtle",
    "bokeh_character": "minimal | creamy"
  },
  "post_processing": {
    "cleanup": "minor dust removal",
    "authenticity": "retain crafting marks",
    "grain": "none | subtle"
  },
  "output": {
    "aspect_ratio": "3:2 | 4:3 | 16:9",
    "resolution": "high",
    "crop_safety": "no tight crops",
    "file_style": "photorealistic | painterly"
  }
}

THIS IS NOT EXTENSIVE. you can add your own values, for example, if you want 6 different specific items, you can create your own JSON item.

Pro tips: If you don’t know how to, describe it to AI and they’ll help you.

Example: Turning This Into AI Gold

Before (Vague Prompt):

“clay dog on moon”

Press enter or click to view image in full size

After (JSON-Engineered Prompt):

{
  "image_purpose": "still_life",
  "scene": {
    "environment": "studio",
    "background": {
      "material": "sculpted clay backdrop",
      "color_tone": "dark-cool",
      "cleanliness": "clean",
      "color_hex": "#2b2b2b"
    },
    "surface": {
      "material": "matte clay base",
      "tone": "neutral-grey",
      "finish": "matte",
      "color_hex": "#7d7d7d"
    }
  },
  "composition": {
    "framing": "vertical",
    "tiers": 1,
    "arrangement_notes": "dog centered on clay moon with background stars evenly distributed for balance",
    "negative_space": "medium",
    "camera_height": "eye-level"
  },
  "camera": {
    "focal_length_mm": 50,
    "aperture_f": 4.5,
    "lens_type": "prime",
    "focus_strategy": "sharp on dog and moon, slightly blurred background",
    "distance_descriptor": "mid"
  },
  "lighting": {
    "type": "softbox",
    "direction": "upper-left",
    "quality": "diffused",
    "contrast": "medium",
    "color_temperature": "neutral-warm",
    "shadow_behavior": "soft edge"
  },
  "materials_and_texture": {
    "primary_materials": ["unglazed clay"],
    "surface_finish": "unglazed",
    "texture_notes": "visible fingerprints, uneven edges, thumb-pressed moon craters",
    "color_palette": {
      "dog_black": "#2f2f2f",
      "dog_white": "#f5f1e8",
      "moon_grey": "#8d8d8d",
      "stars_yellow": "#f4c542",
      "glasses_grey": "#c3c3c3"
    }
  },
  "mood": ["whimsical", "playful", "tactile"],
  "imperfections": {
    "include": true,
    "types": ["fingerprints", "tool marks", "slight misalignments", "uneven clay edges"]
  }
}

“A horizontal still-life photograph of six unglazed beige clay ceramics on two tiers of light matte wooden shelving against a clean, neutral-warm wall. Natural diffused light from the left casts soft shadows. Shot with a 50mm prime lens at f/4, all subjects in crisp focus. Visible imperfections — uneven rims, subtle finger depressions — enhance the handmade feel.”

Generated Image:

Press enter or click to view image in full size

Vocabulary Cheat Sheet

This is what I use, the art in prompt engineering is that you can create your own categories. When getting started, you can use mine as guide.

When you’re trying to describe an image for creative direction, AI prompts, or photography notes, it helps to break it down into specific categories.

1. Image Purpose

What’s the core intent of the shot?

2. Scene Environment

Where does the shot take place?

3. Background

Material

Color Tone

4. Surface Material

What’s the subject sitting on?

5. Composition & Framing

6. Camera & Lens

Focal length changes perspective:

7. Lighting

Type: Natural, strobe, continuous
Direction: Left, right, back, top
Quality: Diffused, soft, hard

8. Color Grading

Palette keywords like:

"warm beige", "soft teal", "charcoal black", "sunset orange"

9. Materials & Texture

10. Mood

Think emotional tone: calm, earthy, inviting, dramatic, luxurious, mysterious.

11. Imperfections

Small details that add character: uneven rims, faint seam lines, patina scratches.

12. Subject Category

13. Depth of Field

14. Post-Processing

15. Output Ratio

If this helped you see AI prompts in a new way:

AI Image Generator Prompt Engineering
Twitter LinkedIn Facebook

Get AI scheduling insights, product news, and Bay Area community updates delivered to your inbox.

No spam. Unsubscribe anytime.

Published by TEAMCAL AI · More articles

2026 © TEAMCAL AI | Zara AI | StudentSync AI