ocr-and-documents

Use pymupdf to open the PDF and call get_text() on each page, or use pymupdf4llm to convert directly to markdown. For remote PDF URLs, try web_extract first since it handles PDF-to-markdown conversion without local dependencies.

pymupdf is lightweight (~25MB) and instant, handling text-based PDFs, tables, and image extraction. marker-pdf (~3-5GB with PyTorch) adds OCR for scanned documents in 90+ languages, equation recognition, forms, and reading-order detection.

No, pymupdf cannot perform OCR on scanned documents. For scanned PDFs you need marker-pdf, which includes OCR support for over 90 languages, or use web_extract if the document is available at a remote URL.

Use python-docx, which parses the actual document structure rather than relying on OCR. Install it with pip install python-docx. For PPTX files, use the powerpoint skill built on python-pptx.

marker-pdf requires roughly 5GB of free disk space for PyTorch and its models, plus about 2.5GB of models downloaded to ~/.cache/huggingface/ on first use. Run the extract_marker.py script with --check to verify disk space before installing.

pymupdf handles splitting and merging natively with no extra dependencies. Use insert_pdf to copy page ranges into a new document for splitting, or combine multiple source documents into one for merging, then save the result.