Start with a simple prompt and automatically create a complex one to make your images perfect for advertising and photography. A complete workflow for Z-IMAGE that includes an LLM agent that translates short, weak prompts into complex, complete ones. A powerful upscaler to reach absurd resolutions without losing quality.

The agent was written to create images that are perfect for advertising and photographic purposes, with particular attention to lights, apertures and exposure times.

Two files are released:

1) Turbo version

2) Pro version (slower)

All nodes install themselves on ComfyUi; simply insert the .json file into your ComfyUi window.

This work with Z-image Model

https://huggingface.co/Tongyi-MAI/Z-Image-Turbo

Here the tutorial:

https://www.youtube.com/watch?v=ujFVUO8Pa-o

Combinatorial Workflow in Flux