Is it possible to run LlamaParse locally without involving LlamaCloud? I see questions posted on the youTube tutorial on if it can be on Prem but not answered. If so, is there a sample notebook?

Did anyone attend the talk yesterday about parsing and metadata tagging by Deasie and Jerry? I’ve been wondering how the tagging is done in the database. In one of the slides he showed that many tags are applied to different parts of a document. How are the tags associated with the specific location in the document? Was the document broken up into chunks and stored individually for each tag? How are the hierarchical relationships among the tags maintained?

I am going though the video "Build Agents from Scratch". At the end, the LLM reasoned that the tool output has satisfied the question and then move to the state is_done. I feel there is a leap of faith in that reasoning: how did it determine the output is enough? In case it did not like the tool output, will it go into an infinite loop?