The v7 version utilizes the brand-new FLUX RF Inversion sampling technology, combined with DeepSeek to accurately reverse-prompt keywords. The similarity of reference image redrawing exceeds 90%, with poses, lighting, and clothing maintaining extremely high consistency, comparable to commercial photoshoot quality!


*Known issue: There is a low probability of image generation failure; please try again!

*If the result is unsatisfactory, modify the random seed until you get a satisfactory result~


v7.0:

Adopts FLUX RF Inversion sampling, DeepSeek reverse-prompt keywords, and highly controls the Flux model.

Simply find a photoshoot reference image on Xiaohongshu, then combine it with your own photo to quickly generate a set of studio-level photoshoots unique to you!


It is recommended to run the workflow online on RunningHub; FLUX is too memory-intensive, and it’s impossible to run without a 4090...

If you have any other issues, feel free to leave a comment, and we will fix them as soon as possible. The workflow will be updated frequently~