LLM itself does not have the ability to interpret PDFs (although there are models that like do…), so I think it depends on whether the interface has a feature that handles PDFs. If you modify that, it should be able to handle other document files as well. Which interface?