About XTORX

A torrent meta-search engine: one search across every major public index, running since 2011.

What it does

You search once. XTORX takes those words to every index in its catalogue at the same time, each in the shape that index expects, and answers with a single page: one entry per index.

There is no crawler, no database and no server doing work on your behalf: the whole search runs inside your own browser. That is why it answers instantly, and why nothing you type is kept anywhere on this side.

What it does not do

It hosts no torrent files, no magnet links and no copies of anything.
It runs no tracker and joins no swarm.
It carries no advertising. The only third party on the page is Google Analytics, which stores nothing until you agree to it.
It never contacts the indexes itself. Your browser does, only when you click.

Privacy

Because there is no backend, nothing is logged on this side, and no search history is kept in your browser. The one third party is Google Analytics. It is loaded with storage denied, so it sets no cookie and keeps no identifier until the bar at the bottom of the page is accepted; refusing is one click in the same place, and the answer can be changed at any time from the Cookies link in the footer. That answer is the only thing this site keeps in your browser.

Analytics is told what you search for. The words go to it as a search event, so the site can see which searches are common, and they are sent on their own rather than pasted into the address of every page: the page a visit reports is just /search/, with no query attached. Nothing sent alongside identifies you. The indexes you then open learn nothing at all, because outbound links carry rel="noopener noreferrer", so the index you open is not told which page sent you or what you typed here.

Open content

The open-content archives that close every result page carry material which is explicitly free to copy and redistribute: Academic Torrents and Linux Tracker. BitTorrent was built for exactly this, and it is still the best way to move a Linux image or a large research dataset.

The 2011 version

XTORX started as a handful of PHP files: a form that POSTed to search.php, which answered with a fixed list of indexes to try. It worked for a decade. It also echoed the raw search term straight into the HTML, which was a cross-site scripting hole, and by the end the host had been compromised and the index page was carrying an injected dropper.

The rebuild keeps the idea and the green wordmark, and drops everything else: no PHP, no trackers, no advertising SDK, and a results page that nothing you type can turn into markup. The old files are kept in the repository, disinfected, as a record.

Indexes that did not survive

Torrent indexes come and go. These were on XTORX at some point and are now gone: BTJunkie, isoHunt, Fenopy, Btmon, ScrapeTorrent, Seedpeer, SUMOTorrent, Torrentbit, TorrentReactor, MNova, Veoble, SkyTorrents, ETTV, ilCorsaroNero, RARBG, Torrentz2, TorrentFunk, BitSearch, Solid Torrents, GloTorrents.