Get (almost) original messages from google group archives. Your data is yours.
Switch branches/tags
Clone or download
Latest commit 657d49b Aug 15, 2018

Build Status

Table of contents


This is a Bash-4 script to download all original messages from a Google group archive. Private groups require you to load cookies from file.


The script requires bash-4, sort, wget, sed, awk.

Make the script executable with chmod 755 and put them in your path (e.g, /usr/local/bin/.)

The script may not work on Windows environment as reported in


The first run

The first run

# export _ORG=""  # only if you are using Gsuite
export _GROUP="mygroup"       # specify your group

./ -sh              # first run for testing
./ -sh >    # save your script
bash                  # downloading mbox files

Update your local archive thanks to RSS feed

After you have an archive from the first run you only need to add the latest messages as shown in the feed. You can do that with -rss option and the additional _RSS_NUM environment variable:

export _RSS_NUM=50            # (optional. See Tips & Tricks.)
./ -rss > # using rss feed for updating

It's useful to follow this way frequently to update your local archive.

Private group or Group hosted by an organization

To download messages from private group or group hosted by your organization, you need to provide cookies in legacy format.

  1. Export cookies for google domains from your browser and save them as file. Please use a Netscape format. A sample cookie file is as below FALSE / FALSE 1461209169 G2_ADLT  1458617169721       TRUE  / FALSE 1474428370 NID      77=a-very-long-string      TRUE  / FALSE 1490153162 __cfduid a-hash-of-some-thing

    (G2_ADLT available when you access some adult-contents group.)

    When you have the file, please open it and remove all #HttpOnly_ strings. See also

  2. Specify your cookie file by _WGET_OPTIONS:

     export _WGET_OPTIONS="--load-cookies /your/path/my_cookies.txt --keep-session-cookies"

    Now every hidden group can be downloaded :)

The hook

If you want to execute a hook command after a mbox file is downloaded, you can do as below.

  1. Prepare a Bash script file that contains a definition of __wget_hook command. The first argument is to specify an output filename, and the second argument is to specify an URL. For example, here is simple hook

     # $1: output file
     # $2: url (
     __wget_hook() {
       if [[ "$(stat -c %b "$1")" == 0 ]]; then
         echo >&2 ":: Warning: empty output '$1'"

    In this example, the hook will check if the output file is empty, and send a warning to the standard error device.

  2. Set your environment variable _HOOK_FILE which should be the path to your file. For example,

     export _GROUP=archlinuxvn
     export _HOOK_FILE=$HOME/bin/

    Now the hook file will be loaded in your future output script.

What to do with your local archive

The downloaded messages are found under $_GROUP/mbox/*.

They are in RFC 822 format (possibly with obfuscated email addresses) and they can be converted to mbox format easily before being imported to your email clients (Thunderbird, claws-mail, etc.)

You can also use mhonarc ultility to convert the downloaded to HTML files.

See also

Rescan the whole local archive

Sometimes you may need to rescan / redownload all messages. This can be done by removing all temporary files

rm -fv $_GROUP/threads/t.*    # this is a must
rm -fv $_GROUP/msgs/m.*       # see also Tips & Tricks

or you can use _FORCE option:

_FORCE="true" ./ -sh

Another option is to delete all files under $_GROUP/ directory.

Known problems

  1. This script may not recover emails from public groups. When you use valid cookies, you may see the original emails if you are a manager of the group. See also
  2. When cookies are used, the original emails may be recovered and you must filter them before making your archive public.

For script hackers

Please skip this section unless your really know to work with Bash and shells.

  1. If you clean your files (as below), you may notice that it will be very slow when re-downloading all files. You may consider to use the -rss option instead. This option will fetch data from a rss link.

    It's recommmeded to use the -rss option for daily update. By default, the number of items is 50. You can change it by the _RSS_NUM variable. However, don't use a very big number, because Google will ignore that.

  2. Because Topics is a FIFO list, you only need to remove the last file. The script will re-download the last item, and if there is a new page, that page will be fetched.

     ls $_GROUP/msgs/m.* \
     | sed -e 's#\.[0-9]\+$##g' \
     | sort -u \
     | while read f; do
         last_item="$f.$( \
           ls $f.* \
           | sed -e 's#^.*\.\([0-9]\+\)#\1#g' \
           | sort -n \
           | tail -1 \
         echo $last_item;
  3. The list of threads is a LIFO list. If you want to rescan your list, you will need to delete all files under $_D_OUTPUT/threads/

  4. You can set the time for mbox output files, as below

     ls $_GROUP/mbox/m.* \
     | while read FILE; do \
         date="$( \
           grep ^Date: $FILE\
           | head -1\
           | sed -e 's#^Date: ##g' \
         touch -d "$date" $FILE;

    This will be very useful, for example, when you want to use the mbox files with mhonarc.


This work is released under the terms of a MIT license.


This script is written by Anh K. Huynh.

He wrote this script because he couldn't resolve the problem by using nodejs, phantomjs, Watir.

New web technology just makes life harder, doesn't it?