
lixorn/deconstructed-duotone-poster
ResearchThis page was translated automatically. Technical details may differ from the original. View the original
This is an image stylization skill for ChatGPT / Codex. It first identifies the subject in the image, then breaks it down into a set of flat graphics to create editorial posters with two-tone, soft light, paper texture, and film graininess.
Overview
This is an image stylization skill for ChatGPT / Codex. It first identifies the subject in the image, then breaks it down into a set of flat graphics to create editorial posters with two-tone, soft light, paper texture, and film graininess. - Transform the photo into a 3:4 vertical nine-grid poster. - Transform the vertical image into a 3:4 vertical quadruptych poster. - Transform the horizontal photo into a 4:3 horizontal six-panel poster. - Transform the horizontal image into a 4:3 horizontal quadruptych poster. - If no photos are provided, you can also generate directly based on the text theme. - You can add a designated subject to the original image. - Fixed on off-white paper, with another theme color you can choose. - Automatically add two lines of small text and minimalist icons related to the main image. Each grid does not draw the same photo repeatedly. Skill selects outlines, movements, materials, and local features, then reexpresses them with different flat graphics. The final image looks more like a printed poster than a photo filter. All four layouts retain a significant amount of blank space. The top and left and right sides are at equal distance, and each grid is a standard square. Quadtychs are suitable for images with a clear subject and a more restrained image. - Large flat graphics, not directly copying the original photo. - Moderate detail, without colored pencils, ribbon lines, or realistic texture. - Light off-white paper background with a theme color. - Graphic edges with a slight hand-printed feel. - Soft light and halo are more pronounced, but the main graphic remains clear. - Features natural paper grain and fine grain. - Only two lines of small text and one to five main icons at the bottom. Download or clone this repository and install the complete deconstructed-duotone-poster folder as a custom skill. Don't just copy the SKILL.md, because layout references and other rule files also need to be retained. If you only want to see the generated prompt, you can request "Prompt only." If you only want to analyze styles, you can request "Analyze only, not generate." The theme color can be directly named as light blue, brick red, forest green, or purple, or can provide hexadecimal color values. The default paper base is fixed in a lighter off-white #F7F1E3.
Overview (machine translated)
English README
This is an image stylization skill for ChatGPT / Codex. It first identifies the subject in the image, then breaks it down into a set of flat graphics to create editorial posters with two-tone, soft light, paper texture, and film graininess.
Example of the effect
What can it do?
- Transform the photo into a 3:4 vertical nine-grid poster.
- Transform vertical images into 3:4 vertical quadtych posters.
- Transform landscape photos into 4:3 landscape six-grid posters.
- Transform the horizontal image into a 4:3 horizontal quadtych poster.
- If you don’t provide photos, you can also generate them directly based on the text theme.
- You can add a specified subject to the original image.
- Off-white paper base fixed, with another theme color you can choose.
- Automatically adds two lines of small text and minimalist icons related to the main image.
Each grid does not draw the same photo repeatedly. Skill selects outlines, movements, materials, and local features, then reexpresses them with different flat graphics. The final image looks more like a printed poster than a photo filter.
Four Layouts
| Layout | Format | Content | Suitable |
|---|---|---|---|
| Nine-Grid | 3:4 Vertical Layout | 3×3 Grid | Vertical Chart, Square Chart, Text Theme |
| Vertical Quartet | 3:4 Vertical | 1×4 Grid | A more concise vertical diagram is needed |
| Six-Grid | 4:3 Landscape | 3×2 Grid | Landscape Image, Wide Scene |
| Horizontal Quartet | 4:3 Landscape | 4×1 Grid | A more concise horizontal diagram is needed |
All four layouts retain a significant amount of blank space. The top and left and right sides are at equal distance, and each grid is a standard square. Quadtychs are suitable for images with a clear subject and a more restrained image.
Visualization Features
- Large flat shapes, do not copy the original photo.
- Moderate detail, without colored pencils, ribbon cables, or realistic texture.
- Light off-white paper background with a theme color.
- The edges of the graphic have a slight hand-printed feel.
- Soft light and halos are more obvious, but the main body of the graphic remains clear.
- Features natural paper grain and fine particles.
- The bottom has only two lines of small text and one to five main icons.
Installation
Download or clone this repository and install the full deconstructed-duotone-poster folder as a custom skill. Don’t just copy SKILL.md, because layout references and other rule files also need to be retained.
Local Codex can use:
How to use
When using it, you only need to provide images and theme colors:
Specify the horizontal six-grid grid:
No images provided, just write the topic:
Add a new subject to the original image:
When processing multiple images at once, you can specify colors separately:
Three main models
| Pattern | What You Need to Provide | Results |
|---|---|---|
| Image mode | One or more images | Extract the subject from each image and redesign it |
| Text Theme Mode | A short topic | Create images directly based on text |
| Image plus subject mode | Original image and subject to add | Keep the original image theme while adding a new subject |
This page was translated automatically. Technical details may differ from the original.
Recommended Tools
Try a different keyword or remove a filter.
Install
npx skillfish add lixorn/deconstructed-duotone-poster