GPT Image 2/2.5 Prompt Gallery is a prompt library and CLI tool for OpenAI image generation and editing with GPT Image 2 and 2.5.
The tool addresses the challenge of crafting effective prompts for image generation by providing a curated gallery of numbered prompt entries organized across categories, alongside reference images and task-specific templates. It works by offering both a command-line interface for direct API interaction and two reusable agent skills that can be integrated into Claude Code, Codex, and other skill-capable agent runtimes. The approach emphasizes using clear visual references—sketches, layouts, or existing images—rather than lengthy text descriptions, allowing users to attach reference materials via the image editing API.
Developers working with GPT Image 2 or 2.5 should choose this tool if they need a starting point for prompt engineering or want to integrate image generation into agentic workflows. It suits projects requiring research figures, presentation graphics, or iterative image editing where reference-based prompting is more effective than text alone. The gallery provides building blocks for vector-style diagrams, icons, and layouts useful for PowerPoint and design work. The CLI accepts PNG, WebP, and JPG reference images and supports explicit edit constraints, making it practical for workflows that begin with exported PDF pages or existing visual materials.
The project maintains an active prompt gallery with entries organized across multiple categories and includes documentation in both English and Chinese. Development activity shows ongoing refinement of the collection and API support, with the tool regularly updated to reflect capabilities of current image generation models.
