[K0balt]

Can multimodal llms read the pdf file format to extract text components as well as graphical ones? Because that would seem to me to be the best way to go.