Ymmiah/glowmint ? reverse-engineered prompt

Build me a modern AI photo editor web app called GlowMint.

I want users to be able to upload photos, switch between editing an existing image and generating a new one from text, and get a clean polished interface that feels fast and easy to use. The app should connect to the Gemini API, support smart prompt improvement for simple ideas, and let people choose common aspect ratios like square and widescreen when generating images.

Please also include useful editing tools like magic erase, magic replace, style remix, 3D photo effect, and a way to batch process a few images at once. I want undo and redo so people can safely experiment, plus a helpful error log with reference IDs and a retry option if the API fails. If it helps, look up the current docs online and use sensible defaults so the app feels complete and production ready.