# When to Use Each of the 21 Industrial Prompt Templates in GPT-Image-2

> Discover when to use each of the 21 industrial prompt templates from freestylefly/awesome-gpt-image-2. Optimize your creative tasks from UI mockups to architectural visuals with these powerful tools.

- Repository: [苍何/awesome-gpt-image-2](https://github.com/freestylefly/awesome-gpt-image-2)
- Tags: deep-dive
- Published: 2026-09-13

---

**The freestylefly/awesome-gpt-image-2 repository provides 21 industrial-grade prompt templates organized by visual-output category, with each template optimized for specific creative tasks ranging from UI mockups to architectural visualizations.**

The **2I (Industrial-Level) Prompt Templates** deliver structured generation schemas for GPT-Image-2. Located in [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md) and [`data/style-library.json`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/data/style-library.json), these templates exist in both human-readable plain-text and machine-readable JSON formats. According to the repository's [`README.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/README.md), this dual-format architecture supports everything from manual prompt crafting to automated agent workflows through the GPT-Image-2 Style Library skill.

## Understanding the 2I Template System

The repository defines **industrial prompt templates** as structured instruction sets that specify `type`, `style`, `content`, and `constraints` for image generation. Each template targets a specific visual domain, ensuring the model receives precise, context-rich directions rather than ambiguous descriptions.

According to [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md), templates follow two implementation patterns:

- **Plain-text "common" versions**: Editable text files for manual copy-paste workflows
- **JSON "advanced" versions**: Schema-validated objects for programmatic use via [`agents/skills/gpt-image-2-style-library/SKILL.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/agents/skills/gpt-image-2-style-library/SKILL.md)

## Digital Product and Interface Design

### UI and Interfaces (tpl-ui)

Use **`tpl-ui`** when generating high-fidelity mockups for applications, websites, or dashboards. This template excels at defining concrete layout structures including navigation hierarchies, card arrangements, and tab systems while specifying target platforms (iOS, Android, Web).

The template enforces **component-level thinking**, requiring definitions for sidebar navigation, main content areas, and interactive elements. For a finance dashboard implementation:

```js
await skill.run(`
{
  "type": "UI Screenshot",
  "platform": "Web",
  "product": "Finance Dashboard",
  "layout": "Sidebar navigation + main data table",
  "style": {
    "theme": "Light",
    "primary_color": "Blue",
    "typography": "Sans-serif"
  },
  "content": {
    "header": "Today's Overview",
    "cards": [
      {"title": "Revenue", "data": "$1.2M"},
      {"title": "Expenses", "data": "$800K"}
    ]
  },
  "constraints": "High fidelity, 16:9 aspect ratio"
}
`);

```

Alternatively, use the plain-text version from [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md) for manual editing:

```text

# Plain-text UI prompt (tpl-ui)

为金融仪表板生成一张Web界面图。
核心功能：实时数据、过滤器、导出按钮。
视觉风格：极简，主色蓝，强调色橙。
布局：左侧导航栏 + 主内容区。
输出：高保真UI截图，文字清晰可读，比例16:9。

```

### Product and E-commerce (tpl-product)

Deploy **`tpl-product`** for commercial product photography and catalog layouts. This template structures prompts around **product selling points**, packaging details, and detail-page compositions. Use it when showcasing single products with highlight callouts and clean commercial aesthetics.

## Visual Marketing and Brand Identity

### Poster and Typography (tpl-poster)

Select **`tpl-poster`** for event, product, or cinematic posters where **bold headline systems** and striking composition dominate. This template prioritizes text prominence and visual impact over data density, making it ideal when mood and message hierarchy outweigh informational complexity.

### Brand and Logos (tpl-brand)

Use **`tpl-brand`** when defining or refreshing visual identity systems. This template generates **logos, color palettes, typography specimens, and usage guidelines**. It ensures consistency across brand touchpoints by structuring prompts around identity constraints rather than single-image outputs.

### Infographic and Information (tpl-infographic)

Choose **`tpl-infographic`** for data-rich visualizations including flowcharts, comparison charts, and timeline graphics. This template excels when conveying structured information with **clear visual connectors** and hierarchical data modules.

The JSON structure enforces scientific-atlas aesthetics and layout discipline:

```js
await skill.run(`
{
  "type": "Infographic",
  "topic": "Renewable Energy Mix",
  "audience": "Policy Makers",
  "structure": {
    "title_area": "全球可再生能源分布",
    "layout": "Vertical flowchart",
    "modules": [
      {"title": "太阳能", "icon": "sun", "text": "占比40%"},
      {"title": "风能", "icon": "wind", "text": "占比30%"},
      {"title": "水能", "icon": "water", "text": "占比20%"},
      {"title": "生物质", "icon": "leaf", "text": "占比10%"}
    ]
  },
  "style": {
    "aesthetic": "Scientific atlas",
    "colors": "Low saturation, color-coded flows",
    "background": "Light paper texture"
  },
  "constraints": "No cyberpunk, strict layout"
}
`);

```

## Photorealistic and Artistic Content

### Photography and Realism (tpl-photo)

Apply **`tpl-photo`** for photographic-style outputs requiring **lens characterization, lighting setups, and realistic textures**. This template simulates camera hardware and environmental lighting to generate photorealistic portraits, street scenes, and object photography.

### Illustration and Art (tpl-illustration)

Use **`tpl-illustration`** for stylized artwork, concept art, or comic-style outputs. This template focuses on **brushwork techniques, material simulation, and art historical styles** rather than photographic accuracy.

### Architecture and Spaces (tpl-architecture)

Deploy **`tpl-architecture`** for architectural visualizations and interior design concepts. This template controls **perspective angles, material specifications, and indoor/outdoor lighting scenarios** to render realistic building scenes and spatial layouts.

## Character and Narrative Development

### Characters and People (tpl-character)

Select **`tpl-character`** for consistent character design across multiple poses or scenes. This template generates **pose sheets, concept references, and stylized portraits** while maintaining visual consistency in clothing, features, and proportions.

### Scene and Storytelling (tpl-scene)

Use **`tpl-scene`** for narrative-driven visuals including storyboards and cinematic concepts. This template structures prompts around **emotional pacing, action sequences, and world-building elements** where composition drives the narrative.

### History and Classical Chinese Themes (tpl-history)

Choose **`tpl-history`** when producing historically accurate Chinese scrolls, period costumes, or ancient settings. This template encodes **dynasty-specific visual languages**, traditional clothing details, and scroll-style narrative structures.

## Document and Specialized Formats

### Documents and Publishing (tpl-document)

Apply **`tpl-document`** for magazine spreads, reports, and editorial layouts. This template enforces **page systems, table of contents structures, and typographic hierarchy rules** where document structure dominates visual design.

### Other Use Cases (tpl-other)

Use **`tpl-other`** as a catch-all for experimental workflows, mixed-media compositions, or hybrid scenarios that do not fit standard categories. This template provides maximum flexibility when specialized constraints would otherwise restrict creative exploration.

## Programmatic Implementation

The **GPT-Image-2 Style Library** skill defined in [`agents/skills/gpt-image-2-style-library/SKILL.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/agents/skills/gpt-image-2-style-library/SKILL.md) consumes the JSON schemas stored in [`data/style-library.json`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/data/style-library.json). For API integration, reference [`src/apimartClient.js`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/src/apimartClient.js) to understand how the repository communicates with generation endpoints.

When building agent workflows, import the JSON templates directly to ensure schema validation. The `skill.run()` method accepts validated template objects and manages the translation to API-compatible prompts.

## Summary

- **Industrial prompt templates** in awesome-gpt-image-2 provide structured, category-specific schemas for GPT-Image-2 generation
- The repository maintains **dual-format templates**: plain-text for manual editing and JSON for programmatic agents
- **UI, Product, and Brand templates** handle commercial digital asset creation with layout and identity constraints
- **Infographic and Document templates** enforce information hierarchy and typographic rules for data-heavy outputs
- **Photo, Illustration, and Architecture templates** control technical parameters like lighting, materials, and perspective
- **Character and Scene templates** maintain consistency across narrative sequences and pose variations
- Implementation references include [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md), [`data/style-library.json`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/data/style-library.json), and the Style Library skill documentation

## Frequently Asked Questions

### What is the difference between plain-text and JSON industrial prompt templates?

Plain-text templates in [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md) provide human-readable prompt structures optimized for manual copying and editing. JSON templates in [`data/style-library.json`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/data/style-library.json) offer machine-readable schemas with typed fields for `type`, `style`, `content`, and `constraints`, enabling programmatic validation through the GPT-Image-2 Style Library skill.

### How do I integrate these templates into my application?

Import the JSON schemas from [`data/style-library.json`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/data/style-library.json) and invoke them through the Style Library skill interface documented in [`agents/skills/gpt-image-2-style-library/SKILL.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/agents/skills/gpt-image-2-style-library/SKILL.md). For direct API calls, reference [`src/apimartClient.js`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/src/apimartClient.js) to understand the HTTP client implementation and authentication patterns required for the generation endpoint.

### Which industrial prompt template works best for brand identity design?

Use **`tpl-brand`** for comprehensive identity work including logos and color systems. For single marketing assets like event announcements, combine **`tpl-poster`** for layout impact with **`tpl-brand`** constraints to ensure typography and color consistency across deliverables.

### Can I combine multiple templates for complex creative workflows?

While each template targets specific output categories, the **`tpl-other`** category accommodates hybrid workflows. For structured combinations, manually merge JSON schema elements from multiple templates, ensuring constraint fields do not conflict. The plain-text versions allow more flexible concatenation for experimental compositions.