desktop-gpt-image-2/gpt-image-2-desktop — reverse-engineered prompt

Build me a simple desktop app for Windows and Mac that lets people use GPT Image 2 like a real creative tool, not a browser tab. I want it to feel fast and easy, open the app, type a prompt, optionally drop in a reference image, hit Generate, and get up to 8 consistent images at 2K quality. It should be good at things people actually want, like posters with readable text, UI mockups that look real, character continuity across variations, image edits on a selected region, and the option to upscale when needed.

Please make it feel private and local. Save prompts, outputs, prompt versions, collections, styles, and reusable characters on the user’s machine with no tracking. Let people search their library, reuse past prompts, export PNG JPG and WebP, and use handy size presets for social posts, wallpapers, and similar formats. If it fits, include a simple cost view and an API request inspector for advanced users. Look up current docs online if you need to.