How to Create a Multi-Style Signature Selection Poster with a 6-Card Grid

You can create a multi-style signature selection poster by defining a JSON layout with six card objects and leveraging the CSS grid definition at src/styles.css#L1368 to render them in a responsive 6-column layout.

The awesome-gpt-image-2 repository provides a JSON-driven prompt system for generating AI-powered posters with customizable grid layouts. By combining a specific CSS grid configuration with a structured JSON payload, you can produce signature selection posters that display six distinct visual styles in a single cohesive image.

Understanding the Grid Architecture

The rendering system relies on a clear separation between the visual layout (CSS) and the content structure (JSON).

The CSS Foundation in src/styles.css

At line 1368 of src/styles.css, the repository defines the six-column grid rule that powers the layout:

grid-template-columns: repeat(6, minmax(0, 1fr));

(src/styles.css#L1368)

This CSS rule ensures that the poster container distributes six cards evenly across the width, creating the foundation for your signature selection grid.

The JSON Layout Structure

Each card in the grid is defined as a photo object within the JSON layout. These objects support individual style arguments including signature text, decorative doodles, and custom messages. The backend generate-image API processes this JSON by merging it with templates from docs/templates.md and style definitions from data/style-library.json.

Building the 6-Card JSON Layout

To create your poster, construct a JSON payload that specifies six cards in the layout grid. Each card description uses placeholder syntax that the prompt engine expands using the style library.

Here is a complete, ready-to-use JSON configuration:

{
  "type": "signature-selection-poster",
  "theme": "multi-style",
  "layout": {
    "grid": {
      "columns": 6,
      "cards": [
        { "position": "1", "description": "signature '{argument name=\"signature text\" default=\"Hanashi\"}' with heart doodle" },
        { "position": "2", "description": "signature '{argument name=\"signature text\" default=\"Hanashi\"}' with sparkle doodle" },
        { "position": "3", "description": "signature '{argument name=\"signature text\" default=\"Hanashi\"}' with bunny doodle" },
        { "position": "4", "description": "signature '{argument name=\"signature text\" default=\"Hanashi\"}' with flower doodle" },
        { "position": "5", "description": "signature '{argument name=\"signature text\" default=\"Hanashi\"}' with star doodle" },
        { "position": "6", "description": "signature '{argument name=\"signature text\" default=\"Hanashi\"}' with message '{argument name=\"message text\" default=\"Thanks for the support!\"}'" }
      ]
    }
  },
  "style": "graffiti-sketch",
  "background": "white",
  "size": "9:16"
}

Key configuration fields:

• layout.grid.columns: Maps directly to the CSS repeat(6, ...) rule and tells the backend to expect six cards.
• cards[].description: Contains placeholder-rich strings using the {argument name="..."} syntax defined in data/style-library.json.
• style: Selects a visual preset from the style library (e.g., "graffiti-sketch").
• background and size: Control the final image dimensions and backdrop.

Integrating with the Prompt Engine

The prompt engine merges your JSON layout with the style library (data/style-library.json) and template files (docs/templates.md) to build the final AI prompt. When processing your request, the engine expands placeholders like {argument name="signature text"} into concrete text values and constructs a single comprehensive prompt for the image model.

The frontend component in src/main.jsx receives the generated image URL and renders it inside a <div class="poster-grid"> container, which automatically applies the six-column CSS grid defined in the stylesheet.

Generating the Poster via API

Send your JSON payload to the /api/generate-image endpoint to generate the poster. Here is a Node.js implementation:

import fetch from 'node-fetch';

const payload = {
  "type": "signature-selection-poster",
  "theme": "multi-style",
  "layout": {
    "grid": {
      "columns": 6,
      "cards": [
        { "position": "1", "description": "signature '{argument name=\"signature text\" default=\"Hanashi\"}' with heart doodle" },
        { "position": "2", "description": "signature '{argument name=\"signature text\" default=\"Hanashi\"}' with sparkle doodle" },
        { "position": "3", "description": "signature '{argument name=\"signature text\" default=\"Hanashi\"}' with bunny doodle" },
        { "position": "4", "description": "signature '{argument name=\"signature text\" default=\"Hanashi\"}' with flower doodle" },
        { "position": "5", "description": "signature '{argument name=\"signature text\" default=\"Hanashi\"}' with star doodle" },
        { "position": "6", "description": "signature '{argument name=\"signature text\" default=\"Hanashi\"}' with message '{argument name=\"message text\" default=\"Thanks for the support!\"}'" }
      ]
    }
  },
  "style": "graffiti-sketch",
  "background": "white",
  "size": "9:16"
};

const res = await fetch('https://your-domain.com/api/generate-image', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify(payload)
});

const data = await res.json();
console.log('Poster URL:', data.imageUrl);

The API returns a JSON object containing the imageUrl of your generated multi-style signature selection poster.

Summary

• The 6-card grid is defined in src/styles.css at line 1368 using grid-template-columns: repeat(6, minmax(0, 1fr)).
• JSON layout objects control the content of each card through the layout.grid.cards array, supporting custom signatures and doodles.
• Placeholder syntax ({argument name="..."}) in card descriptions links to definitions in data/style-library.json.
• The /api/generate-image endpoint processes the JSON and returns a rendered poster URL.
• The frontend in src/main.jsx displays the result using the CSS grid container class poster-grid.

Frequently Asked Questions

What file controls the 6-column grid layout in awesome-gpt-image-2?

The CSS rule that creates the six-column layout is located at line 1368 in src/styles.css. It uses grid-template-columns: repeat(6, minmax(0, 1fr)) to distribute six cards evenly across the poster width.

How do I customize the signature text for each card in the grid?

Each card's description field in the JSON layout uses placeholder syntax like {argument name="signature text" default="Hanashi"}. You can modify the default value in the JSON or configure the style library in data/style-library.json to accept dynamic inputs via the API request.

Can I change the number of cards from 6 to a different number?

Yes. To display a different number of cards, update the layout.grid.columns value in your JSON payload and modify the corresponding CSS rule in src/styles.css to use repeat(N, minmax(0, 1fr)) where N is your desired card count. The JSON layout and CSS grid work independently but must align numerically.

Where are the visual styles defined for the signature posters?

Visual styles are stored in data/style-library.json and referenced by the style field in your JSON payload (e.g., "graffiti-sketch"). The prompt engine merges these style definitions with your layout specifications and the templates in docs/templates.md to generate the final AI prompt.

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