I came across this gem of a Python package:
It converts varying types of documents into markdown and it does so very good (it is even managing to describe PowerPoint Diagrams created with Smart Art accurately without the need to send an image to an LLM).
Currently trying to integrate this into my AI extension as I think in the context of preparing data for RAG this is a game changer - it’s running in Dev Mode but having issues with the build (thats what below topic is about):
I think for advanced users that can set up their own Python envs this might be a need way to get word docs etc. into KNIME easily.