3 ms·

This is an interesting approach! Why not offload PDF extraction to other frameorks that apply OCR pdf -> .md

• blackmanta 1y ago
  I may explore this when I implement the vectordb implementation I started.