I have been building GPT Image Lab, a visual prompt library for gpt-image-2.
The goal is simple: make GPT Image 2 prompts easier to browse, evaluate, copy, and reuse. Instead of keeping prompts as isolated text snippets, every item is paired with a preview image, category, source metadata, and a path back to the original example.
Current snapshot, using the latest available data from 2026-04-26:
- Website library: 2,985 prompt/template items
- Prompts: 1,060
- Templates: 1,925
- Newly added on 2026-04-26: +703
- New prompts: +443
- New templates: +260
- Open-source GitHub collection: 500 curated prompts
Links:
The website supports browsing by image and generating images from prompts in one click. The GitHub repo provides a smaller curated set with preview images and machine-readable exports.
Why I built it
Great gpt-image-2 prompt examples are scattered across social posts, prompt sites, Reddit threads, and demos. The image might be useful, but the prompt is hard to find. Or the prompt is available, but you cannot quickly judge whether it is worth reusing.
So I organized the project around three layers:
- A website for visual browsing and one-click image generation.
- A GitHub repo with 500 curated, image-backed prompts.
- JSON and CSV exports for developers who want to analyze or reuse the data.
The repository currently includes:
- 500 copy-ready GPT Image 2 prompts
- Preview images
- Categories such as Infographic, Poster, UI Mockup, Product Visual, 3D Scene, and more
- Source attribution
data/gpt-image-2-prompts.jsondata/gpt-image-2-prompts.csv
A few examples
1. Historical miniature grid
Useful for dense educational visuals, timelines, and grid-based scene generation.
Prompt excerpt:
Create a 16x16 grid with equal cells, each showing a different culture, civilization, or region-specific scene from a different time period in human history...
2. Isometric city weather scene
Useful for city visuals, weather explainers, and stylized 3D scenes.
Prompt excerpt:
Present a clear, 45 degree top-down view of a vertical isometric miniature 3D cartoon scene, highlighting iconic landmarks centered in the composition...
3. Packaging dieline to 3D box
Useful for packaging design, ecommerce mockups, and brand presentation images.
Prompt excerpt:
Assemble the dieline into a flawless 3D box with accurate panels, clean folds, undistorted type, and artwork preserved exactly...
4. Mobile app onboarding screen
Useful for UI mockups, app concepts, and product design exploration.
Prompt excerpt:
Create a vertical mobile onboarding screen for a fictional app called NESTING. Headline: WELCOME TO NESTING...
5. Product cutout with transparent background
Useful for ecommerce product images and marketing assets.
Prompt excerpt:
Extract the product from the input image. Output: transparent background, crisp silhouette, clean edges, no halos, no fringing...
6. Technical infographic
Useful for explainers, technical education, and social media diagrams.
Prompt excerpt:
Style: tech-focused, futuristic, minimal but visually engaging. Include short captions, avoid clutter...
7. City food map illustration
Useful for travel guides, local content, and editorial maps.
Prompt excerpt:
Create a hand-drawn city food map themed around Chengdu, using a simplified illustrated city map with landmarks and food spots...
8. Recipe infographic
Useful for food content, recipe cards, and step-by-step visual tutorials.
Prompt excerpt:
Create step-by-step recipe infographic for creamy garlic mushroom pasta, top-down view, minimal style on white background...
9. Photoshop shortcuts infographic
Useful for quick-reference cards and design education.
Prompt excerpt:
Create a clean and modern infographic showing the most essential Photoshop shortcuts, well organized and easy to read.
10. Data science and machine learning panorama
Useful for learning roadmaps, technical posters, and course visuals.
Prompt excerpt:
Generate an extremely complex, information-dense panoramic infographic about data science and machine learning.
What I want to improve next
Next, I want to keep improving:
- Better categories and filters
- More prompt templates
- Better examples for ecommerce, posters, infographics, UI mockups, and thumbnails
- Cleaner contribution flow for the GitHub repo
- More structured metadata for downstream analysis
If you are experimenting with gpt-image-2, this might be useful as a reference library:










Top comments (0)