awesome-gpt-image-2

This repository is a large, community-curated library of text prompts for OpenAI's GPT Image 2 model, which is an AI system that generates images from written descriptions. If you tell the model what you want to see, it produces an image matching your description, and the quality of the description has a big effect on the quality of the result. This collection exists to save time by providing ready-made, tested prompts that other users have found effective.

The library currently contains over 2,000 prompts and is updated daily. Each entry comes with a preview of the image it produces, so you can see what the output looks like before trying it. The prompts are organized by category across two dimensions: use case (such as product marketing images, YouTube thumbnails, social media posts, comic storyboards, and game assets) and visual style (such as watercolor, pixel art, 3D render, anime, oil painting, or cyberpunk). Subject categories cover portraits, food, architecture, landscapes, and more.

The README itself is a long list of prompts with image previews, but the maintainers also run a separate web gallery on their site that displays the same content in a grid layout with full-text search, category filters, and a button to generate an image directly from a prompt without copying it manually. The GitHub repository is the source of truth; the website is a more convenient way to browse.

The project is available in 16 languages, with separate README files for Chinese, Japanese, Korean, Spanish, German, French, and others. Contributions from the community are welcome via pull requests. The content is licensed under Creative Commons Attribution 4.0, meaning it can be reused freely with credit. The full README is longer than what was shown.