ocr-and-documents

To extract text from a scanned PDF using OCR, this skill applies marker-pdf, which is layout-aware and supports over 90 languages. It automatically selects the appropriate OCR-based extractor for scanned documents to convert images into readable text.

The best way to extract text from a PDF URL is using the web_extract method. This skill prioritizes web_extract as the first step to fetch remote PDFs and convert them directly into text or Markdown format.

Yes, pymupdf works for text extraction from native text-based PDFs. This skill uses pymupdf specifically for local text-based PDF files, while reserving marker-pdf for OCR-based scans to ensure accurate document processing.

You can extract text from DOCX and PPTX files, but this skill provides compatibility guidance rather than direct parsing. It recommends using dedicated external tools for Word and slides to achieve better extraction results.

You should use marker-pdf instead of pymupdf for document processing when handling OCR-based scans. The skill selects pymupdf for local text-based PDFs and switches to marker-pdf for scanned documents requiring layout-aware OCR.