Neon AI Image Workflow Infographic

{
  "type": "Dark Technology Workflow Infographic",
  "style": "The presentation slides feature a clean, futuristic aesthetic, a gradient background from black to deep navy blue, accented with neon cyan light effects, slender cyan connecting arrows, modern UI cards with rounded corners, soft shadows, high-contrast white text, and sophisticated startup brand aesthetics.",
  "canvas": {
    "aspect_ratio": "16:9",
    "composition": "The title block is left-aligned, with modular charts on the right and bottom."
  },
  "branding": {
    "topline": "Codex × GPT-5.5 × ChatGPT Images 2.0",
    "title": "GPT-Image-Harness",
    "subtitle": "Transform image generation from a \"card-drawing\" process into a production workflow."
  },
  "layout": {
    "sections": [
      {
        "title": "Left-side function list",
        "position": "Left sidebar",
        "count": 3,
        "labels": [
          "Organization materials",
          "Multiplexing instructions",
          "Accumulated correction"
        ]
      },
      {
        "title": "Reference materials (input)",
        "position": "Upper middle left",
        "count": 2,
        "labels": [
          "Character design illustrations",
          "Past comic pages"
        ]
      },
      {
        "title": "Prompt words/rules",
        "position": "Upper middle to right",
        "count": 1,
        "labels": [
          "# Output Rules"
        ]
      },
      {
        "title": "Codex / CLI",
        "position": "Left side of the center",
        "count": 3,
        "labels": [
          "Environment setup",
          "CLI / Skill Execution",
          "Automation scripts"
        ]
      },
      {
        "title": "GPT-Image-Harness",
        "position": "center",
        "count": 6,
        "labels": [
          "assets/",
          "reference/",
          "output/",
          "prompts/",
          "skills/",
          "AGENTS.md",
          "SPEC.md"
        ]
      },
      {
        "title": "Output (comic page)",
        "position": "Right side of the center",
        "count": 1,
        "labels": [
          "Comic pages"
        ]
      },
      {
        "title": "Bottom operation",
        "position": "bottom row",
        "count": 2,
        "labels": [
          "Manual adjustment and fine-tuning",
          "Generate re-correction prompts using ChatGPT"
        ]
      }
    ]
  },
  "left_panel": {
    "headline_scale": "extra large",
    "title_weight": "Bold and overweight",
    "decorative_line": "The short horizontal cyan line below the title block",
    "feature_items": [
      {
        "icon": "folder outline glowing cyan",
        "title": "Organization materials",
        "description": "Unified management of role settings and past pages"
      },
      {
        "icon": "terminal window outline glowing cyan",
        "title": "Multiplexing instructions",
        "description": "Template the prompts and rules"
      },
      {
        "icon": "bar chart outline glowing cyan",
        "title": "Accumulated correction",
        "description": "Retain test logs to improve reproducibility"
      }
    ]
  },
  "diagram": {
    "nodes": [
      {
        "kind": "input card",
        "title": "Reference materials (input)",
        "visual": "Two thumbnail areas: one a half-body portrait of an anime character (with a softly blurred face), and the other a collage of black and white comic book pages."
      },
      {
        "kind": "rule card",
        "title": "Prompt words/rules",
        "visual": "List of bullet points in the code style panel",
        "body_text": [
          "# Output Rules",
          "- Maintain role consistency",
          "- Configure storyboards according to specified requirements",
          "- Background density should be consistent with the previous page.",
          "- No mechanical details omitted",
          "- Does not include speech bubbles and text"
        ]
      },
      {
        "kind": "tool card",
        "title": "Codex / CLI",
        "visual": "A terminal window with the prompt >_ and three bullet points"
      },
      {
        "kind": "repository card",
        "title": "GPT-Image-Harness",
        "visual": "Large central file tree panel with a cyan glowing border"
      },
      {
        "kind": "output card",
        "title": "Output (comic page)",
        "visual": "Two comic book page thumbnails side by side"
      },
      {
        "kind": "action card",
        "title": "Manual adjustment and fine-tuning",
        "visual": "round pencil icon",
        "description": "The addition of speech bubbles, text, and balance adjustments were all done manually."
      },
      {
        "kind": "action card",
        "title": "Generate re-correction prompts using ChatGPT",
        "visual": "Speech bubble icon with three dots",
        "description": "Clearly define the direction of correction, repeat the process, and then output again."
      }
    ],
    "connectors": "A cyan glowing arrow connects the input, rules, and CLI to the central project card; arrows point from the project card to the output and down to the two bottom operation cards; a dashed cyan feedback arrow loops from the output back to the lower right operation area."
  },
  "typography": {
    "language": "Simplified Chinese",
    "title_color": "White",
    "accent_color": "bright cyan",
    "body_color": "light gray-white"
  },
  "lighting": "Soft neon edge light around cards and icons",
  "quality": "Clear vector-style infographics, demo-quality, precise alignment, and no watermark."
}