You paste a link to a file, our server downloads it and uploads it to FileDitch, and you get an ordinary FileDitch link. Nothing goes through your own connection, and you can close the tab: the transfer runs on our server. The status page is the way back to it, so keep its address.
Direct links only: an http:// or https:// address that starts a download when you open it. Not supported: torrents and magnet links; pages that show a download button (Mega, Google Drive, MediaFire, YouTube, Telegram and the like); streaming playlists (m3u8, mpd); and links that need a login, cookies or a captcha, or that are tied to your own IP address. A Dropbox share link does work: we ask for the file instead of the preview page.
We look at the link first, then download it. The first look takes a few seconds and reads little more than the headers, so a dead or unsupported link fails at once instead of after a wait in the queue. A link that only works a single time is used up by that first look: give us one that can be fetched more than once.
150 GB, the same as on FileDitch. If the site does not say how big the file is we still try, but only 2 of those run at the same time.
The ones FileDitch refuses: programs, installers, scripts and web pages (.exe, .apk, .js, .html and friends), judged by name and by content. Links to FileDitch itself are refused too: that file is already there.
One server does all of this, with 1 TB of disk and a 1 Gbit/s line, and every file has to fit on it before it goes to FileDitch. When the queue is long or the disk is full, new links wait or the form closes for a while. Per IP address: 20 links per 10 minutes and 150 per day, 10 in progress, 2 transferring at the same time, 300 GB per 24 hours. If you share an address with other people (a VPN, a campus), you share these too.
About as long as downloading the file and uploading it again, plus the queue. The front page shows how long people have been waiting lately; the status page shows speed and time left. A source that is too slow to finish in 12 h is dropped, and a link that waits in the queue for more than 4 h is given up on, because by then it has usually expired.
Your IP address, the link you gave us, the file name, your browser's user-agent, the time, and the FileDitch link that came out. FileDitch itself only sees our server as the uploader, so this record is what ties an upload to whoever asked for it. We keep it without a time limit and hand it over when law enforcement asks, like FileDitch does. The site you download from sees our server's address, not yours. Anyone who has the address of a status page can see that upload.
A list your browser keeps for you (localStorage, up to 50 entries, 30 days). It is never sent to us, it does not follow you to another device, and clearing the site data for this site, or the button under the list, empties it.
It is an ordinary FileDitch file and follows the usual FileDitch rule: a file that goes 45 days without a single access can be deleted when the space is needed. The details are in the FileDitch FAQ. We keep nothing of the file here once it is uploaded.
Our requests come from one server and say who they are: Mozilla/5.0 (compatible; FileDitchRemoteUpload/1.0; +https://remote.fileditch.com/faq) — one first look, then one download per link somebody gave us, a few at a time per site, no crawling. If you want your host excluded, tell us and we will block it.