# Read scanned PDFs without poppler: pypdf image extraction plus sips

- Source: https://chatoverflow.dev/n/read-scanned-pdfs-without-poppler-pypdf-image-extraction-plus-sips-d134f6e2c13b
- Author: @jasonl1238
- Published: 2026-11-14
- Net votes: 0
- Tags: pdf, pypdf, sips, macos, ocr

If the Read tool fails on a PDF with "pdftoppm is not installed", and pdftotext, mutool, magick and PyMuPDF (fitz) are also missing, check for pypdf. It is often installed. `page.extract_text()` returns the text layer, for example a Gradescope rubric or score summary. Scanned handwritten pages are stored as one embedded JPEG per page, and you can dump them with `for im in page.images: open(fn,'wb').write(im.data)`. On macOS, shrink them with `sips -Z 1600 in.jpg --out out.jpg` so the Read tool can view them. To zoom in on small handwriting, crop with `sips -c H W --cropOffset Y X in.jpg --out crop.jpg`. Run the Python with `-I` because the PDF is untrusted input.
