How to Back up a Tumblr Blog
This will be a long post.
Big thank you to @afairmaiden for doing so much of the legwork on this topic. Some of these instructions are copied from her verbatim.
Now, we all know that tumblr has an export function that theoretially allows you to export the contents of your blog. However, this function has several problems including no progress bar (such that it appears to hang for 30+ hours) and when you do finally download the gargantuan file, the blog posts cannot be browsed in any way resembling the original blog structure, searched by tag, etc.
What we found is a tool built for website archiving/mirroring called httrack. Obviously this is a big project when considering a large tumblr blog, but there are some ways to help keep it manageable. Details under the cut.
How to download your blog with HTTrack:
Website here
You will need:
A reliable computer and a good internet connection.
Time and space. For around 40,000 posts, expect 48 hours and 40GB. 6000 posts ≈ 10 hours, 12GB. If possible, test this on a small blog before jumping into a major project. There is an option to stop and continue an interrupted download later, but this may or may not actually resume where it left off. Keep in mind that Tumblr is a highly dynamic website with things changing all the time (notes, icons, pages being updated with every post, etc).
A custom theme. It doesn't have to be pretty, but it does need to be functional. That said, there are a few things you may want to make sure are in your theme before starting to archive:
the drop down meatball menu on posts with the date they were posted
tags visible on your theme, visible from your blog's main page
no icon images on posts/notes (They may be small, but keep in mind there are thousands of them, so if nothing else, they'll take up time. Instructions on how to exclude them below.)
Limitations: This will not save your liked or private posts, or messages. Poll results also may not show up.
What to expect from HTTrack:
HTTrack will mirror your blog locally by creating a series of linked HTML files that you can browse with your browser even if tumblr were to entirely go down. The link structure mimics the site structure, so you should be able to browse your own blog as if you had typed in the url of your custom theme into the browser. Some elements may not appear or load, and much of the following instructions are dedicated to making sure that you download the right images without downloading too many unnecessary images.
There will be a fair bit of redundancy as it will save:
individual posts pages for all your tags, such as tagged/me etc (If you tend to write a lot in your tags, you may want to save time and space by skipping this option. Instructions below.)
the day folder (if you have the meatball menu)
regular blog pages (page/1 etc)
How it works: HTTrack will be going through your url and saving the contents of every sub directory. In your file explorer this will look like a series of nested folders.
How to Start
Download and run HTTrack.
In your file directory, create an overarching folder for the project in some drive with a lot of space.
Start a new project. Select this folder in HTTrack as the save location for your project. Name your project.
For the url, enter https://[blogname].tumblr.com. Without the https:// you'll get a robots.txt error and it won't save anything.
Settings:
Open settings. Under "scan rules":
Check the box for filetypes .gif etc. Make sure the box for .zip etc. is unchecked. Check the box for .mov etc.
Under "limits":
Change the max speed to between 100,000 - 250,000. The reason this needs to be limited is because you could accidentally DDOS the website you are downloading. Do not DDOS tumblr.
Change the link limit to maybe 200,000-300,000 for a cutoff on a large blog, according to @afairmaiden. This limit is to prevent you from accidentally having a project that goes on infinitely due to redundancy or due to getting misdirected and suddenly trying to download the entirety of wikipedia.
Go through the other tabs. Check the box that says "Get HTML first". Uncheck "find every link". Uncheck "get linked non-html files". If you don't want to download literally the entire internet. Check "save all items in cache as well as HTML". Check "disconnect when finished".
Go back to Scan Rules.
There will be a large text box. In this box we place a sort of blacklist and whitelist for filetypes.
Paste the following text into that box.
+*.mp4 +*.gifv -*x-callback-url* -*/sharer/* -*/amp -*tumblr.com/image* -*/photoset_iframe/*
Optional:
-*/tagged/* (if you don't want to save pages for all your tags.)
-*/post/* (if you don't want to save each post individually. not recommended if you have readmores that redirect to individual posts.)
-*/day/* (if you don't feel it's necessary to search by date)
Optional but recommended:
-*/s64x64u*.jpg -*tumblr_*_64.jpg -*avatar_*_64.jpg -*/s16x16u*.jpg -*tumblr_*_16*.jpg -*avatar_*_16.jpg -*/s64x64u*.gif -*tumblr_*_64.gif -*avatar_*_64.gif -*/s16x16u*.gif -*tumblr_*_16.gif -*avatar_*_16.gif
This will prevent the downloading of icons/avatars, which tend to be extremely redundant as each image downloads a separate time for each appearance.
Many icons are in .pnj format and therefore won't download unless you add the extension (+*.pnj), so you may be able to whitelist the URLs for your and your friends' icons. (Honestly, editing your theme to remove icons from your notes may be the simpler solution here.)
You should now be ready to start.
Make sure your computer doesn't overheat during the extremely long download process.
Pages tend to be among the last things to save. If you have infinite scroll on, your first page (index.html) may not have a link to page 2, but your pages will be in the folder.
Shortly after your pages are done, you may see the link progress start over. This may be to check that everything is complete. At this point, it should be safe to click cancel if you want to stop, but you run the risk of more stuff being missing. You will need to wait a few minutes for pending transfers to be competed.
Once you're done, you'll want to check for: Files without an extension.
Start with your pages folder, sort items by file type, and look for ones that are simply listed as "file" rather than HTML. Add the appropriate extension (in this case, .html) and check to see if it works. (This may cause links to this page to appear broken.)
Next, sort by file size and check for 0B files. HTMLs will appear as a blank page. Delete these. Empty folders. View files as large icons to find these quickly.
If possible, make a backup copy of your project file and folder, especially if you have a fairly complete download and you want to update it.
Finally, turn off your computer and let it rest.
