Topic: Is there any way to search based on image code?

Posted under e621 Tools and Applications

If for example, I have the full image code of an image. Like this:

5e28ad15a779183f4575ab5d50cd9d69

This code appears when you download an image off the site as the default name of the image.

The actual image is here:

post #450707

Only problem is. When you go to download it says this:

http://static1.e621proxy.ru/data/5e/28/5e28ad15a779183f4575ab5d50cd9d69.jpg

Note how the code is the same:

Code from download: 5e28ad15a779183f4575ab5d50cd9d69

Code from site when you click "download": 5e28ad15a779183f4575ab5d50cd9d69

Only problem is this part of the link:

http://static1.e621proxy.ru/data[b]/5e/28/[/b]5e28ad15a779183f4575ab5d50cd9d69.jpg

This chunk of code: /5e/28/

Cannot be guessed. Because it's different for every page. With other codes like: /6f/dc/ or /43/87/

Updated

well, it's interesting to note the first four digits in the long hash are the same as the two, two-digit chunks seperated by slashes.

look again at .../5e/28/5e28...

and for the following image, it's the same: .../38/b8/38b8...

post #1216881

Updated by anonymous

The "image code" is MD5 hash. This way you can always check that image is the same by checking hash and comparing it to filename.
Like this: https://puu.sh/vQ50y/84ac4501a7.png

You can search using MD5 hash with md5:hash

So for this image, you search with md5:5e28ad15a779183f4575ab5d50cd9d69
(I think inkbunny is one of the only other sites allowing searching with hash?)

This is also one of the main reasons why you shuold NEVER alter the file from original before upload unless you really have to (too large res, ugoira conversion, etc.). Because hash is easiest and most reliable way of telling that two files are the same, this way upload cancels if you try to upload same image and it's easy to find image here even if you found it elsewhere.

Also bit unrelated, but I have been trying to make wiki page with this from other sites, having all sites URL format is something that would be amazing to have, because e.g. deviantart, you can actually take single ID from post URL to get into submission URL, but something like FA has useless data like UNIX timestamp instead.
https://e621.net/wiki/show/howto:sites_and_sources

Updated by anonymous

Mario69 said:
[...]

i have not found evidence that DeviantART, Pixiv, or FurAffinity allows MD5 searching.

Updated by anonymous

Sorry for late replies!

But it appears that doing:

md5:370781d8bd71bec5cc812ef3523ad6e8

Into the search bar does work. So thanks!

Updated by anonymous

  • 1