Stripping the link annotations from a PDF before it goes into Zotero would kill the highlighting, since Zotero’s boxes are drawn from the PDF’s actual link/annotation objects (and sometimes from auto-detected URI text). A few ways to do it, roughly in order of ease:
1. If you have Acrobat Pro Edit → “Remove Hidden Information” (a.k.a. Redaction/Sanitize) can strip all link annotations in one pass, or use Edit → Link → Remove All Links if available in your version.
2. Free/cross-platform: `pikepdf` or `qpdf` (Python) This is the most reliable scriptable route. Link annotations are stored as /Annots with subtype /Link. A short script can strip them:
import pikepdfpdf = pikepdf.open("input.pdf")for page in pdf.pages: if "/Annots" in page: page.Annots = pikepdf.Array([ a for a in page.Annots if a.get("/Subtype") != pikepdf.Name("/Link") ])pdf.save("output_no_links.pdf")
This removes the clickable link annotations specifically, without touching highlights, text, or bookmarks.
3. Free GUI option: PDF-XChange Editor (Windows) or PDFsam Some let you batch-remove all link annotations via a menu option without scripting.
4. Caveat on citation-style “links” Some of what you’re seeing (especially the exponent/citation false positives someone mentioned) may not be true /Link annotations at all — Zotero’s reader may be auto-linkifying plain URL/DOI-looking text in the content stream itself. If stripping /Annots doesn’t fix it, that’s likely what’s happening, and there’s no easy way to un-linkify detected text patterns without altering the underlying text layer (which risks breaking searchability/copy-paste).
If you want, upload a sample PDF and I can run the link-stripping script on it directly and hand back a cleaned copy to test in Zotero.