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:
Website: https://gptimagelab.com/
GitHub: https://github.com/peterRooo/awesome-gpt-image-2-prompts
The website supports browsing by image and generating images from prompts in one click.…