ocr-and-documents

Use pymupdf to open the PDF and call get_text() on each page, or pymupdf4llm for markdown output. It installs in seconds at about 25MB and handles text-based PDFs, tables, images, and metadata without downloading models.

pymupdf is lightweight (~25MB) and instant but cannot OCR scanned pages, equations, or forms. marker-pdf (~3-5GB with PyTorch) handles OCR in 90+ languages, LaTeX equations, and complex layouts at roughly 1-14 seconds per page on CPU.

Scanned PDFs require OCR, which pymupdf does not support. Use marker-pdf, which performs OCR in over 90 languages and outputs markdown; check for about 5GB of free disk space before installing since it downloads PyTorch and models.

Yes. pymupdf's find_tables() detects tables and converts them to pandas DataFrames for basic cases. For high-accuracy table extraction in complex layouts, marker-pdf produces better structured markdown output.

Empty output usually means the PDF is a scanned image with no text layer, so standard parsers find nothing. Switch to an OCR-capable tool like marker-pdf, or if the document has a URL, try web_extract first.

pymupdf handles both natively with no extra dependencies. Use insert_pdf() to copy page ranges into a new document for splitting, or to combine multiple files for merging, then save the result.