skip to navigation
skip to content

ws.rsspull 2.2

Downloads RSS/Atom feeds and converts them to Maildir messages.

rsspull downloads and parses RSS and Atom feeds, converts posts into email messages, and stores them in Maildirs. This way you can read your feeds using an email client (I use claws, so I can access mailing lists, newsgroups and feeds all in one application). The heavy lifting is performed by the awesome feedparser library.


rsspull requires at least Python 2.6 (and won’t work under Python 3 yet). You can install it from PyPI like this:

$ pip install ws.rsspull

You need to create a configuration file in ~/.rsspull/config, like:

target = ~/Maildir/rss
target_type = maildir
logfile = ~/.rsspull/log
workers = 1

For low-volume applications you can also send the emails via SMTP (localhost) instead:

target =
target_type = smtp

List the feeds you want to pull in an OPML file at ~/.rsspull/feeds.opml:

<?xml version="1.0" encoding="utf-8"?>
<opml version="1.1">
    <outline text="tech">
      <outline text="ongoing" xmlUrl="" />
    <outline text="general">
      <outline text="heisec" xmlUrl="" />
      <outline text="trac_example" xmlUrl=";ticket=on&amp;changeset=on&amp;wiki=on&amp;max=10&amp;daysback=90&amp;format=rss" auth="user:password"/>

Notes about the format:

  • You can group your feeds (using nested <outlines>), this does not matter to rsspull.
  • The text attribute is used as the folder name, relative to the maildir setting in ~/.rsspull/config. If you run rsspull on an IMAP server, you can create folder hierarchies by using something like text="" (check you IMAP server documentation how it represents folder hierarchies. The example with dots works for Courier, others might use actual subfolders, and so on).
  • Basic authentication is supported with the auth attribute.
  • You can use file:// URLs.

Then simply run:

$ rsspull

to download the feeds.

You can pass rsspull --confdir /path/to/config to use a different location than ~/.rsspull.


  • Uses If-Modified-Since HTTP headers to avoid downloading a feed that has not changed.

  • Addds a Content-Location header to each message that contains the URL of that post. I bound the following script to [return] in claws to open the current entry in a webbrowser:

    URL=`sed -ne '/^Content-Location/s/.*: //p' $1 | head -n 1`
    if [ -n "$URL" ]; then
        mozilla $URL &> /dev/null
  • Entries are converted to multipart messages, one with the original HTML and one converted into markdown-like plaintext (via html2text), so most of the time opening in a browser isn’t even necessary since you can read the post right in the email client.

  • Can use several worker threads to download and parse feeds to increase performance, since quite some time is spent waiting for downloads to complete (the workers setting in ~/.rsspull/config).

  • Has been in daily usage since 2007, so it definitely Works For Me(tm).


2.2 (2018-03-24)

  • Support sending messages via SMTP as an alternative to writing into Maildir.
  • Add commandline option to set config directory.

2.1 (2018-03-06)

  • Switch from urllib2 to requests.
  • Make Python-3 compatible.

2.0 (2014-02-22)

  • Switch from spawning curl to using urllib2.
  • Package as egg.

1.2 (2013-10-03)

  • Use stdlib xml parser instead of python-xml.

1.1 (2009-05-23)

  • Write messages directly to Maildir instead of using procmail.
  • Send messages as multipart, text and html.

1.0 (2007-03-31)

File Type Py Version Uploaded on Size
ws.rsspull-2.2.tar.gz (md5) Source 2018-03-24 47KB