# What Are the 13 Visual Categories for GPT-Image2 Prompt Templates?

> Explore the 13 visual categories for GPT-Image2 prompt templates. Discover standardized structures for UI design, infographics, and more in this essential guide.

- Repository: [苍何/awesome-gpt-image-2](https://github.com/freestylefly/awesome-gpt-image-2)
- Tags: getting-started
- Published: 2026-09-06

---

**GPT-Image2 prompt templates are organized into 13 visual categories (the I³ taxonomy) spanning UI design, infographics, product photography, character art, and document layouts—each with standardized prompt structures defined in [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md).**

The `freestylefly/awesome-gpt-image-2` repository provides an industrial-grade prompt library containing 393 templates for OpenAI's GPT-Image2 model. These templates follow a strict **I³ visual taxonomy** (Image, Illustration, Information, plus extended sub-groups) that helps users rapidly select the right prompt structure for their generative AI needs.

## The 13 I³ Visual Categories Explained

Each category in the taxonomy is documented in [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md) with an anchor link (`#tpl-` prefix) and includes a description, prompt shape, and runnable example.

### UI & Interface

**Core focus:** Screenshots of apps or website interfaces.

**Prompt shape:** `为[产品类型]生成一张[平台]界面图 …`

```text
为健身 App 生成一张 iOS 界面图，核心功能：计步、心率，视觉风格：极简，主色：蓝色

```

Anchor: `#tpl-ui`

### Infographic

**Core focus:** Structured data visualizations and explanatory graphics.

**Prompt shape:** `生成[主题]信息图 …`

```text
生成老年人日常健康管理指南信息图，包含流程图和图标

```

Anchor: `#tpl-infographic`

### Poster & Layout

**Core focus:** Stand-alone visual posters for campaigns and concepts.

**Prompt shape:** `设计一张[活动]海报 …`

```text
设计一张电影海报，主题：星际旅行，风格：暗黑科幻

```

Anchor: `#tpl-poster`

### Product & E-commerce

**Core focus:** Hero shots and catalog imagery for products.

**Prompt shape:** `生成[商品名]电商主图 …`

```text
生成噪音消除耳机的电商主图，材质：磨砂金属，灯光：侧光

```

Anchor: `#tpl-product`

### Brand & Identity

**Core focus:** Logo designs, color systems, and complete brand-asset packages.

**Prompt shape:** `为[品牌名]设计品牌视觉方案 …`

```text
为 Nova Dynamics 设计品牌视觉方案，关键词：创新、极简、可信

```

Anchor: `#tpl-brand`

### Architecture & Space

**Core focus:** Interior and exterior architectural renders.

**Prompt shape:** `生成[空间类型]设计效果图 …`

```text
生成现代简约客厅的渲染图，材质：木材、玻璃，光线：自然采光

```

Anchor: `#tpl-architecture`

### Photography

**Core focus:** Hyper-realistic photo-style images with camera specifications.

**Prompt shape:** `拍摄主题：[人物/物品] …`

```text
拍摄一张咖啡师擦拭咖啡杯的写实照片，使用 50 mm 镜头、f/1.4

```

Anchor: `#tpl-photography`

### Illustration

**Core focus:** Illustrated art spanning anime, hand-drawn, and flat styles.

**Prompt shape:** `创作[题材]插画 …`

```text
创作一幅巨鲸载着小村庄的幻想插画，风格：水彩

```

Anchor: `#tpl-illustration`

### Character

**Core focus:** Character concept art and pose sheets.

**Prompt shape:** `设计[角色身份]角色设定图 …`

```text
设计一名赛博赏金猎人的角色设定图，装备：霓虹管线外套

```

Anchor: `#tpl-character`

### Scene & Narrative

**Core focus:** Story-driven scene compositions.

**Prompt shape:** `生成[故事主题]场景图 …`

```text
生成古代印章碎裂瞬间的场景图，灯光：中心光束

```

Anchor: `#tpl-scene`

### Historical & Period

**Core focus:** Period-accurate historical and traditional Chinese ("古风") scenes.

**Prompt shape:** `生成[朝代]题材画面 …`

```text
生成唐代夜景的宫廷画面，人物：贵妃，色调：暖金

```

Anchor: `#tpl-historical`

### Document & Publication

**Core focus:** Layouts for magazines, menus, and reports.

**Prompt shape:** `制作[文档类型] …`

```text
制作时尚杂志双页展开，使用 3-栏网格，配色：单色红

```

Anchor: `#tpl-document`

### Other Applications

**Core focus:** Miscellaneous use cases including concept boards and R&D materials.

**Prompt shape:** `任务目标：[你要生成的内容类型] …`

```text
为概念产品研发生成完整拆解板，包含迭代步骤与材料说明

```

Anchor: `#tpl-other`

## Template Structure and Syntax Patterns

Each visual category enforces a **standardized prompt syntax** to ensure consistent outputs. The repository defines these patterns in [`data/style-library.json`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/data/style-library.json), which provides color palettes, layout presets, and tone modifiers referenced across templates.

### Complete Template Examples by Category

```text

# UI & Interface Template

为[产品类型]生成一张[平台]界面图。
核心功能：[功能点A]、[功能点B]、[功能点C]。
视觉风格：[极简/科技/拟物]，主色[颜色]，强调色[颜色]。

# Infographic Template

生成[主题]信息图，目标读者为[人群]。
结构：标题区 + [3-5]个模块（每模块含图标、短标题、说明）。

# Poster & Layout Template

设计一张[活动]海报，主题为[主题词]。
主视觉：[主体元素]，标题文案：[标题]，副标题：[副标题]。

# Product & E-commerce Template

生成[商品名]电商主图，卖点为[卖点1]、[卖点2]。
场景：[纯色棚拍/生活方式场景]，镜头：[特写/半身]。

```

## Key Repository Files for Visual Categories

Understanding the file structure helps navigate the 13 visual categories programmatically:

| File | Purpose |
|------|---------|
| [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md) | Master documentation with all 13 I³ categories, anchored sections (`#tpl-ui` through `#tpl-other`), and full template specifications |
| [`data/style-library.json`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/data/style-library.json) | JSON-encoded style definitions referenced by templates across categories |
| [`src/apimartClient.js`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/src/apimartClient.js) | Example client implementation showing how to select a template by category and submit to GPT-Image2 |
| [`README.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/README.md) | Repository overview with I³ taxonomy summary |

## Summary

- The **13 visual categories for GPT-Image2 prompt templates** form the I³ taxonomy: UI & Interface, Infographic, Poster & Layout, Product & E-commerce, Brand & Identity, Architecture & Space, Photography, Illustration, Character, Scene & Narrative, Historical & Period, Document & Publication, and Other Applications.
- Each category lives in [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md) with a `#tpl-` anchor for direct linking.
- Templates follow **strict syntactic patterns** with bracketed variables for consistent prompt generation.
- The [`data/style-library.json`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/data/style-library.json) file centralizes visual style definitions used across categories.
- The [`src/apimartClient.js`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/src/apimartClient.js) reference implementation demonstrates category-based template selection.

## Frequently Asked Questions

### How do I reference a specific visual category in my code?

Use the anchor prefix from [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md). Each of the 13 visual categories has a predictable anchor: `#tpl-ui`, `#tpl-infographic`, `#tpl-poster`, `#tpl-product`, `#tpl-brand`, `#tpl-architecture`, `#tpl-photography`, `#tpl-illustration`, `#tpl-character`, `#tpl-scene`, `#tpl-historical`, `#tpl-document`, and `#tpl-other`. The [`src/apimartClient.js`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/src/apimartClient.js) file shows how to map these anchors to API calls.

### What does "I³" stand for in the GPT-Image2 taxonomy?

The I³ framework represents three core pillars—**Image** (photorealistic renders), **Illustration** (stylized artwork), and **Information** (data-driven visuals)—plus extended sub-groups that expand into the full 13 visual categories. This structure originated in the repository's design documentation to help users quickly identify the right prompt family for their generative task.

### Where are the color palettes and style presets defined?

Visual style definitions live in [`data/style-library.json`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/data/style-library.json). This JSON file encodes color palettes, layout presets, lighting configurations, and tone modifiers that templates across all 13 visual categories reference. Keeping styles centralized ensures consistency when prompts from different categories appear in the same project.

### Can I mix templates from different visual categories?

Yes. The repository architecture encourages compositional workflows. The [`src/apimartClient.js`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/src/apimartClient.js) example demonstrates submitting multiple template IDs in sequence, and the [`style-library.json`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/style-library.json) shared reference system ensures visual coherence across category boundaries. For brand consistency, use the same style tokens (defined in [`data/style-library.json`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/data/style-library.json)) across UI, Product, and Brand category templates.