139

I want to get all the files for a given website at archive.org. Reasons might include:

  • the original author did not archived his own website and it is now offline, I want to make a public cache from it
  • I am the original author of some website and lost some content. I want to recover it
  • ...

How do I do that ?

Taking into consideration that the archive.org wayback machine is very special: webpage links are not pointing to the archive itself, but to a web page that might no longer be there. JavaScript is used client-side to update the links, but a trick like a recursive wget won't work.

4
  • 28
    I've came accross the same issue and I've coded a gem. To install: gem install wayback_machine_downloader. Run wayback_machine_downloader with the base url of the website you want to retrieve as a parameter: wayback_machine_downloader http://example.comMore information: github.com/hartator/wayback_machine_downloader
    – Hartator
    Aug 10, 2015 at 6:32
  • 3
    A step by step help for windows users (win8.1 64bit for me) new to Ruby, here is what I did to make it works : 1) I installed rubyinstaller.org/downloads then run the "rubyinstaller-2.2.3-x64.exe" 2) downloaded the zip file github.com/hartator/wayback-machine-downloader/archive/… 3) unzip the zip in my computer 4) search in windows start menu for "Start command prompt with Ruby" (to be continued)
    – Erb
    Oct 2, 2015 at 7:40
  • 3
    5) follow the instructions of github.com/hartator/wayback_machine_downloader (e;.g: copy paste this "gem install wayback_machine_downloader" into the prompt. Hit enter and it will install the program...then follow "Usage" guidelines). 6) once your website captured you will find the files into C:\Users\YOURusername\websites
    – Erb
    Oct 2, 2015 at 7:40
  • Answers to superuser.com/questions/14403/… also apply to this question Feb 5 at 16:15

5 Answers 5

102

I tried different ways to download a site and finally I found the wayback machine downloader - which was built by Hartator (so all credits go to him, please), but I simply did not notice his comment to the question. To save you time, I decided to add the wayback_machine_downloader gem as a separate answer here.

The site at http://www.archiveteam.org/index.php?title=Restoring lists these ways to download from archive.org:

  • Wayback Machine Downloader, small tool in Ruby to download any website from the Wayback Machine. Free and open-source. My choice!
  • Warrick - Main site seems down.
  • Wayback downloaders - a service that will download your site from the Wayback Machine and even add a plugin for WordPress. Not free.
4
28

This can be done using a bash shell script combined with wget.

The idea is to use some of the URL features of the wayback machine:

  • http://web.archive.org/web/*/http://domain/* will list all saved pages from http://domain/ recursively. It can be used to construct an index of pages to download and avoid heuristics to detect links in webpages. For each link, there is also the date of the first version and the last version.
  • http://web.archive.org/web/YYYYMMDDhhmmss*/http://domain/page will list all version of http://domain/page for year YYYY. Within that page, specific links to versions can be found (with exact timestamp)
  • http://web.archive.org/web/YYYYMMDDhhmmssid_/http://domain/page will return the unmodified page http://domain/page at the given timestamp. Notice the id_ token.

These are the basics to build a script to download everything from a given domain.

6
10

You can do this easily with wget.

wget -rc --accept-regex '.*ROOT.*' START

Where ROOT is the root URL of the website and START is the starting URL. For example:

wget -rc --accept-regex '.*http://www.math.niu.edu/~rusin/known-math/.*' http://web.archive.org/web/20150415082949fw_/http://www.math.niu.edu/~rusin/known-math/

Note that you should bypass the Web archive's wrapping frame for START URL. In most browsers, you can right-click on the page and select "Show Only This Frame".

3
  • 1
    This was greatly useful and super simple! Thanks! I noticed that even though the START URL was a specific Wayback version, it pulled every date of the archive. This may be circumvented by adjusting the ROOT URL, however. Mar 31, 2020 at 15:32
  • Update to my previous comment: The resources in the site may be spread across various archive dates, so the command did not pull all the versions of the archive. You will need to then merge these back into a single folder and clean up the HTML. Mar 31, 2020 at 16:43
  • 1
    this actually worked for me, although I removed the --accept-regex part, otherwise not the whole page was downloaded
    – FurloSK
    Apr 9, 2021 at 7:58
5

There is a tool specifically designed for this purpose, Warrick: https://code.google.com/p/warrick/

It's based on the Memento protocol.

1
  • 4
    As far as I managed to use this (in May 2017), it just recovers what archive.is holds, and pretty much ignores what is at archive.org; it also tries to get documents and images from the Google/Yahoo caches but utterly fails. Warrick has been cloned several times on GitHub since Google Code shut down, maybe there are some better versions there. May 31, 2017 at 16:41
0

I was able to do this using Windows Powershell.

  • go to wayback machine and type your domain
  • click URLS
  • copy/paste all the urls into a text file (like VS CODE). you might repeat this because wayback only shows 50 at a time
  • using search and replace in VS CODE change all the lines to look like this
Invoke-RestMethod -uri "https://web.archive.org/web/20200918112956id_/http://example.com/images/foobar.jpg" -outfile "images/foobar.jpg"
  • using REGEX search/repl is helpful, for instance change pattern example.com/(.*) to example.com/$1" -outfile "$1"

The number 20200918112956 is DateTime. It doesn't matter very much what you put here, because WayBack will automatically redirect to a valid entry.

  • Save the text file as GETIT.ps1 in a directory like c:\stuff
  • create all the directories you need such as c:\stuff\images
  • open powershell, cd c:\stuff and execute the script.
  • you might need to disable security, see link

Not the answer you're looking for? Browse other questions tagged or ask your own question.