Removing idle uploads

Hello again!

I am going to keep this update short, but try to justify/give as much background as I can. Beginning July 6th, 2026 and applying retroactively, files uploaded anonymously that have not had a download/hit in over 2 years will be removed from Catbox.

Catbox has been running for a little over 11 years now, and as such has accumulated what I can only call cruft. Prior to the introduction of Litterbox (and even still to this day really), many people use Catbox as a dumping ground for their script outputs, junk data shares, etc, without considering that this is a free/community funded service. This has lead to an intense amount of data (currently 66 TB) that has not been accessed in over a year, potentially more. Normally the ebb and flow of Lain would keep the idle/active files balanced, with increasing usage, and the exponential increased cost of hard drives due to AI datacenter expansion, I’m forced between a rock and a hard place: put thousands of dollars of my personal money towards additional storage space, or begin culling. And if you think the “thousands of dollars” is an exaggeration, below is the standard hard drive I purchased September 2025 for Catbox servers:

image

And here it is today, July 2nd, 2026 almost a year later:

image

And, since I would really not like data loss, I purchase 2 of these drives for parity. Unfortunately for everyone involved, I don’t have $1,968 to drop every 3 months (about how long it takes Catbox to fill up 20 TB)! Before the prices for everything sky rocketed, I could handle $800~ every 3 months. Before you say something, yes, I have looked at the refurbished market as well - their prices are similarly dismal.

So, now that the justification is out of the way, I’m sure there will be some blowback for this, as I’ve always positioned Catbox as “a filehost that keeps your files until the heat death of the universe”. Perhaps that was my teenage nativity. For reading this far though, you get the special info: the files that are removed will be going to single-layer (non redunant) storage. They will no longer be accessible on files.catbox.moe. If there is a file that you found on a forum post from years ago that has been culled, feel free to email me, and I can pull the file back into the fold.

It sucks that it’s come to this, but with the ouroboros of money that’s happening in the AI bubble, we can only hope for a very sharp needle to blow it up.

Thanks, -cats

2 weeks ago   ·   67 Notes

67 Notes

  1. tvmblrsillyman reblogged this from qtcatbox
  2. yourcherryvampire reblogged this from qtcatbox
  3. verdanttroopertreaty said: All good things come to an end.. because of AI
  4. redcodi said: I’m shocked this wasn’t already done, it’s very reasonable!
  5. gymnocladus-dioicus said: @jestinjoculators i’ve always felt the same way!! thought i was the only one
  6. cheezeexp reblogged this from qtcatbox
  7. 1420mhz said: @percypanleo > Presumably they’re either using RAID 1 to maximize parity to minimize the chance of data loss Don’t see how RAID 1 minimises the chance of data loss. If you have 12 storage drives and 4 redundancy drives, 16 in total, you may lose any 4 of those, in any combination, and still be able to recover all data. If just two consecutive drives of a RAID 1 array malfunction, you lose all data there.
  8. percypanleo said: @1420mhz Presumably they’re either using RAID 1 to maximize parity to minimize the chance of data loss and/or they’re using a version of RAID that doesn’t support extending or changing the type of array (In which case they can either add an additional two drive RAID 1 array every time more space is required or they could buy 5-8 more drives in order to move all of the data over so that they could then reformat the original array, which would take ages given the sheer volume of data)
  9. 1420mhz said: > And, since I would really not like data loss, I purchase 2 of these drives for parity. You… Know about RAID technology, right? Where data that would, without redundancy, take n disks is mathematically distributed between n+1 disks in such a way that failure of any single one of them can be reverted. n+2, n+3, n+4… are also possible. You could have 12 disks, 10 for data and 2 for redundancy, and failure of any two could be corrected. I also would really like to see what compression algorithm and implementation do you use. Modern ones (e.g. Brotli), shared dictionaries between small files and reuse of repeating data chunks between uploads could shave 10–20% or even more off total requirements compared to a more conventional compression e.g. gzip. > feel free to email me If it isn’t already there, would be nice to have that as a pop-up on pages of removed files. I recently watched a great YouTube video about searching for the first ever SkyBlock map (can recommend), the main trouble was that most (except, I recall, MediaFire?) file sharing resources used by forum people in early 2010s did not stand the test of time, so the map is considered to be lost forever. And yeah, I made a Tumblr account specifically to write this comment.
  10. paledbychoice reblogged this from qtcatbox
  11. ilikecatsanpuppies reblogged this from qtcatbox
  12. jestinjoculators said: Honestly I’m kind of relieved. Sometimes I uploaded files that I realised I didn’t need to upload at all, and I always felt bad I was taking up wasted space.
  13. ren-lv reblogged this from qtcatbox
  14. innerlmnt reblogged this from qtcatbox
  15. surelynotshirley said: 1000% fair. The price increase is absolutely insane. I really hope the bubble pops this year. Dystopian tech.
  16. kumit0 said: this is fair, and donos are not something you could always rely on
Older Post