A community member asked if converting PDFs/PNGs to HTML using OCR would be better for their "rag". The comments suggest that LlamaParse, a free tool for up to 1K pages per day, could be a good option to handle the parsing requirements. Another community member mentioned they were trying to keep things free, as they were broke. Overall, the comments indicate that LlamaParse could be a viable solution, but the original poster's financial situation is a consideration.