> I don't see the word powerpoint anywhere in https://github.com/QwenLM/Qwen-Image-Layered,

The word "powerpoint" is not there, however this text is:

“The following scripts will start a Gradio-based web interface where you can decompose an image and export the layers into a pptx file, where you can edit and move these layers flexibly.”

Oh okay I missed it, sorry. But that’s just using a separate python-pptx package to export the generated list of images to a .pptx file, not something inherent to the model.