ideogram-oss/ideogram4 — reverse-engineered prompt

Reverse engineered prompt

GitHub

Build me a simple local image generator using Ideogram 4. I want to type a normal description or paste the structured JSON prompt format, choose the model version if possible, set basic options like image size, seed, and number of images, then save the results to an outputs folder with useful filenames.

Please handle Hugging Face login and gated model access clearly, with friendly errors if I have not accepted the model terms or if my computer cannot run it. Include a clean command line flow and, if it is easy, a small browser page for entering prompts and seeing the generated images.

Add examples for design posters, typography, multilingual text, layout boxes, and color palettes so I can test the model quickly. Keep setup simple with a README that tells me exactly what to install and how to run it. Look up current docs online if you need to.

Want more depth? Deep Reverse