Skip to content


Subversion checkout URL

You can clone with
Download ZIP
A script to extract the main article text from an arbitrary webpage.
Python Shell
Branch: master

Webarticle2Text - Extracts the main article text from a webpage.


Attempts to locate and extract the largest cluster of text in a webpage. It does this by walking the DOM-tree, identifying all text segments and their depth inside the DOM, appends all text at roughly the same depth, and then returns the chunk with the largest total length.

This approach usually works well with typical news sites where one news article is displayed per URL. This approach usually fails with URLs displaying multiple news blurbs (e.g. news aggregators).


You may need to install the tidylib system package, which you can get on Ubuntu 12.04 using:

sudo apt-get install libtidy-0.99-0

or on Fedora using:

sudo yum install libtidy

Then, simply install the package using pip:

pip install webarticle2text


You can invoke the script either as a Python module:

import webarticle2text
print webarticle2text.extractFromURL("http://some/arbitrary/url")

or as a standalone command line script: http://some/arbitrary/url

Note, to use it from the command line, you'll need to ensure it has execute permission and is located in your PATH. On most platforms, this should automatically be done by


  • 1.0.0 (2008.9.16) Initial public release.
  • 1.2.0 (2011.1.3) Update to support Unicode.
  • 1.2.2 (2011.12.17) Cleaned up installation procedure and documentation and moved to
  • 1.2.3 (2011.12.21) Fixed encoding error when redirecting stdout. e.g. http://some/arbitrary/url > output.txt
  • 1.2.5 (2012.11.5) Added the option to specify user-agent header to use when requesting URLs.
  • 2.0.0 (2014.4.20) Added support for Python 3.2.
Something went wrong with that request. Please try again.