[lxe]

Parsing pdfs or web semantically is really not an easy job, as I found in my own foray into LLM sumamrization.