Skip to content


Subversion checkout URL

You can clone with
Download ZIP
The PDF::Reader library implements a PDF parser conforming as much as possible to the PDF specification from Adobe.
Pull request Compare This branch is 485 commits behind master.

Fetching latest commit…

Cannot retrieve the latest commit at this time

Failed to load latest commit information.


The PDF::Reader library implements a PDF parser conforming as much as possible to the PDF specification from Adobe.

It provides programmatic access to the contents of a PDF file with a high degree of flexibility.

The PDF 1.7 specification is a weighty document and not all aspects are currently supported. I welcome submission of PDF files that exhibit unsupported aspects of the spec to assist with improving our support.

This is primarily a low-level library that should be used as the foundation for higher level functionality - it's not going to render a PDF for you. There are a few exceptions to support very common use cases like extracting text from a page.


The recommended installation method is via Rubygems.

gem install pdf-reader


Begin by creating a PDF::Reader instance that points to a PDF file. Document level information (metadata, page count, bookmarks, etc) is available via this object.

reader ="somefile.pdf")

puts reader.pdf_version
puts reader.metadata
puts reader.page_count can accept an IO stream or a filename. Here's an example with an IO stream:

require 'open-uri'

io     = open('')
reader =

PDF is a page based file format, so most visible information is available via page-based iteration

reader ="somefile.pdf")

reader.pages.each do |page|
  puts page.fonts
  puts page.text
  puts page.raw_content

If you need to access the full program for rendering a page, use the walk() method of PDF::Reader::Page.

class RedGreenBlue
  def set_rgb_color_for_nonstroking(r, g, b)
    puts "R: #{r}, G: #{g}, B: #{b}"

reader   ="somefile.pdf")
page     =
receiver =

For low level access to the objects in a PDF file, use the ObjectHash class. You can build an ObjectHash instance directly:


or via a PDF::Reader instance:

reader  ="somefile.pdf")
puts reader.objects

The second method is preferred to increase the effectiveness of internal caching.

Text Encoding

Internally, text can be stored inside a PDF in various encodings, including zingbats, win-1252, mac roman and a form of Unicode. To avoid confusion, all text will be converted to UTF-8 before it is passed back from PDF::Reader.

Strings that contain binary data (like font blobs) will be marked as such on M17N aware VMs.


There are two key exceptions that you will need to watch out for when processing a PDF file:

MalformedPDFError - The PDF appears to be corrupt in some way. If you believe the file should be valid, or that a corrupt file didn't raise an exception, please forward a copy of the file to the maintainers (preferably via the google group) and we can attempt to improve the code.

UnsupportedFeatureError - The PDF uses a feature that PDF::Reader doesn't currently support. Again, we welcome submissions of PDF files that exhibit these features to help us with future code improvements.

MalformedPDFError has some subclasses if you want to detect finer grained issues. If you don't, 'rescue MalformedPDFError' will catch all the subclassed errors as well.

Any other exceptions should be considered bugs in either PDF::Reader (please report it!).



This library is distributed under the terms of the MIT License. See the included file for more detail.

Mailing List

Any questions or feedback should be sent to the PDF::Reader google group. It's better that any answers be available for others instead of hiding in someone's inbox.


The easiest way to explain how this works in practice is to show some examples. Check out the examples/ directory for a few files.

Known Limitations

Occasionally some text cannot be extracted properly due to the way it has been stored, or the use of invalid bytes. In these cases PDF::Reader will output a little UTF-8 friendly box to indicate an unrecognisable character.


Something went wrong with that request. Please try again.