Need help with pdfannots?
Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

156 Stars 47 Forks Other 61 Commits 6 Opened issues


Extracts and formats text annotations from a PDF file

Services available


Need anything else?

Contributors list

# 344,245
50 commits
# 158,976
4 commits
# 73,140
3 commits
# 388,949
1 commit


This is a script that extracts annotations (highlights, comments, etc.) from a PDF file, and formats them as markdown text. It is intended for use in reviewing conference papers.

At present, the following annotations are supported:

  • Highlights without an attached comment are output first, as "highlights" with just the highlighted text included. Note that these are not typically suitable for use in a review, since they're unlikely to have any meaning to the recipient; they are just meant to serve as a reminder to the reviewer.

  • Highlights with an attached comment, and text annotations (not attached to any particular text/highlight) are output next, as "detailed comments". Typically most comments on a reviewed paper are of this form.

  • Underline, strikeout, and squiggly underline annotations are output last, as "Nits", with or without an attached comment. The intention of this is to easily separate formatting or grammatical corrections from more substantial comments about the content of the document.

For each annotation, the page number is given, along with the associated (highlighted/underlined) text, if any. Additionally, if the documents includes outlines (aka bookmarks) such as those generated by the hyperref package, those are also used to identify to which section in the document the annotation refers.


The intended usage of this tool is as a reviewing aid. I typically review a paper directly in my PDF reader, annotating it liberally as I read it (and using squiggly/underline for minor nits). When I've finished reading the paper, I use this script to convert the annotations to plain text, which I paste into the appropriate section of the review form and then edit accordingly to add a high-level review, correct formatting, fix typos, update or remove inaccurate comments, etc.

See --help
for options and invocation.


At present the script has a number of limitations:

  • pdfminer (the underlying PDF parser) sometimes fails to extract text from PDF files when other converters (e.g. pdftotext) do just fine; this results in annotations with some or all text missing (in which case you'll see

    missing text
  • The output from strikeout annotations is not very meaningful

  • When extracting text, we remove all hyphens that immediately precede a line break and join the adjacent words. This usually produces the best results for Latex output (e.g. "soft-\nware" becomes "software"), but sometimes the hyphen needs to stay (e.g. "memory-\nmapped", which will be extracted as "memorymapped"), and we can't tell the difference.



I'd like to change how the output is formatted.

Some minor changes (e.g.: word wrap, skipping sections) can be accomplished via command-line arguments. See --help

All of the output comes from the

class; more elaborate changes can probably be accomplished there.

I think I got a review generated by this tool...

I hope that it was a constructive review, and that the annotations helped the reviewer give you more detailed feedback so you can improve your paper. This is, after all, just a tool, and it should not be an excuse for reviewer sloppiness. Note that I am not the only user of this script.


Andrew Baumann

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.