What Are the I² Visual Categories in awesome-gpt-image-2? A Complete Field Guide

The freestylefly/awesome-gpt-image-2 repository defines thirteen I² visual categories—such as UI & Interfaces, Charts & Infographics, and History & Classical Chinese Themes—that classify every prompt template and generated image in the gallery.

The awesome-gpt-image-2 project organizes its vast collection of GPT‑Image‑2 prompts and visual examples into a strict taxonomy called the I² visual categories. These categories appear in the repository’s README.md under the “🗂️ Category Overview” section and power both the browsable documentation and the programmatic gpt-image-2-style-library skill.

The 13 I² Visual Categories

Although often referenced informally as “the 12 visual categories,” the official I² taxonomy maintained in README.md and data/style-library.json contains thirteen distinct classes. Twelve cover specific visual disciplines, while the thirteenth acts as a catch‑all for experimental or mixed workflows. Every entry in the style library includes a category field that maps to one of the following values:

1. UI & Interfaces – Mobile apps, websites, dashboards, social media screenshots, and product interface mockups.

2. Charts & Infographics – Structured diagrams, knowledge maps, technical explainers, data visualisations, and information graphics.

3. Posters & Typography – Event posters, book covers, type‑driven visuals, and compositions where layout hierarchy dominates.

4. Products & E‑commerce – Product photography, detail pages, packaging designs, selling‑point highlights, and advertisement layouts.

5. Brand & Logos – Logos, identity systems, brand touchpoints, and campaign‑specific visual assets.

6. Architecture & Spaces – Architectural renders, interior designs, city maps, and spatial concept art.

7. Photography & Realism – Portraits, phone‑style photography, film‑texture emulation, and commercial‑grade realistic imagery.

8. Illustration & Art – Vector or raster illustrations, material experiments, decorative images, and fine‑art styles.

9. Characters & People – Character design sheets, pose references, character cards, and 3‑D toy renders.

10. Scenes & Storytelling – Storyboards, narrative scenes, livestream overlay frames, and world‑building visuals.

11. History & Classical Chinese Themes – Classical scrolls, historical figure portraits, traditional motifs, and poetry‑inspired visuals.

12. Documents & Publishing – White‑papers, technical manuals, encyclopedic plates, and multi‑page publishing layouts.

13. Other Use Cases – Creative experiments, mixed workflows, special one‑off tasks, and practical cases that fall outside the standard taxonomy.

Each category links to an anchored section in docs/gallery.md (for example, #cat‑ui for UI & Interfaces) where the full collection of curated examples resides.

How to Access the Categories Programmatically

The canonical list of categories is stored in data/style‑library.json. Each object in the array contains a category property that matches one of the thirteen I² visual categories above. You can extract the unique taxonomy set in JavaScript as follows:

// Load the style-library JSON and list all visual categories
import styleLibrary from '../data/style-library.json' assert { type: 'json' };

const categories = new Set();
styleLibrary.forEach(entry => {
  // Each entry contains a `category` field such as "UI & Interfaces"
  categories.add(entry.category);
});

console.log('I² visual categories:');
categories.forEach(cat => console.log('- ', cat));

Running this script prints the thirteen category names exactly as defined in the repository’s taxonomy, ensuring your automation stays synchronized with the official classification.

Key Source Files Defining the Taxonomy

Three primary files govern the I² visual category system:

• README.md – Declares the categories in the “🗂️ Category Overview” section and links to gallery anchors.
• docs/gallery.md – Hosts the anchored sections (#cat‑ui, #cat‑infographic, #cat‑poster, etc.) that contain the full case listings for each category.
• data/style‑library.json – Machine‑readable source that powers the project website and the gpt‑image‑2‑style‑library skill; every record includes a category field matching the I² visual categories.

Summary

• The awesome-gpt-image-2 repository uses a formal I² visual taxonomy consisting of thirteen categories, despite common references to “12 categories.”
• Categories range from UI & Interfaces to Documents & Publishing, covering design, photography, illustration, and classical themes.
• The classification system is hard‑coded in README.md, documented in docs/gallery.md, and machine‑readable in data/style-library.json.
• Developers can extract category lists programmatically by reading the category field from the style‑library JSON file.

Frequently Asked Questions

Why does the repository list thirteen visual categories instead of twelve?

The “Other Use Cases” category acts as a thirteenth bucket for experimental prompts and workflows that do not fit the standard twelve disciplines. While the project name and common shorthand sometimes imply a dozen categories, the official I² taxonomy in README.md and data/style-library.json explicitly includes thirteen entries to ensure comprehensive coverage.

How do I browse images for a specific visual category?

Navigate to docs/gallery.md in the repository and use the anchored links provided in the README (for example, #cat‑photography for Photography & Realism). Each anchor jumps to a dedicated section containing curated prompt templates and generated image examples for that category.

What is the relationship between the categories and the gpt-image-2-style-library skill?

The gpt‑image‑2‑style‑library skill consumes data/style-library.json at runtime. When the skill generates or classifies images, it references the category field in each JSON record to apply the correct styling rules, layout constraints, and prompt modifiers associated with that I² visual category.

Can I contribute new prompts to a specific visual category?

Yes. Contributors should add prompt entries to data/style-library.json and ensure the category value matches one of the thirteen official I² visual categories. After updating the JSON, place example outputs in the corresponding anchored section of docs/gallery.md so the visual documentation remains synchronized with the data source.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}