Image input like napkins

[zhoupro]

Image input when select a vision model. napkins can upload image but can't modify multi times. llamacoder can modify multiple times, but no image.

[zhoupro]

coder may be use your code with simple modify.

[Nutlope]

Done!