Workflow that turns MiniMax H3 into a powerful image editing model. It supports multiple reference and can do pose copy too.
Useful tips:
In the detailed_description part of the prompt describe style matching your input image - monochrome sketch, expressive digital illustration, anime screencap - it helps preserving 100% of the original style
Target megapixel size should not be lower than 2.0 MP
Description
This version is mainly focused on posing characters






