pdf

name	pdf
description	Use when tasks involve reading or reviewing PDF files. Extract text directly or do visual checks by rendering pages when layout or suspected missing data.
disable-model-invocation	true

PDF

• Use Poppler tools (check pdftotext or pdftoppm help) to read PDFs quickly
• Use pdftoppm when layout matters or there is non-text relevant data (scanned pages, images, tables, ...)
  • Poppler does not OCR scanned PDFs. Render pages first, then look at them id needed
• If needed, work and generate temporary files under /tmp/pdfs/

Working with PDFs

If you need to generate or do something extra, use uvx to run packages like reportlab, pdfplumber, pypdf.