PrivTracker – Private BitTorrent tracker for everyone
privtracker.com184 points by prvt 2 days ago
184 points by prvt 2 days ago
Worth noting if someone uses this to share private data: Checking the "Private torrent" would tell your and compliant clients to not use DHT and other public methods that can leak the content. But it doesn't stop other non-compliant clients from sharing it on the DHT once they have it, nor does it stop someone from just ticking "Enable DHT" on their side once they have received it (or change "private" to "false" in the torrent file itself).
Obvious to many I'm sure, but maybe best to be explicit about it anyways.
Slightly off-topic but kind of fitting: How does infohash v2 support look like today? It's been available for years, but seemingly most private trackers + most other places seems to still be using v1. What clients are people using today, do those support v2? As far as I know, all modern clients do, so it would be possible to start using v2 exclusively.
Reason for the question is that I'm planning to distribute many large files to the public, and in my experience, BitTorrent works really well for that. Question is if it's enough to just publish v2 infohashes, or I need to publish both v1 and v2.
If you flip the "private" flag it would change the infohash of the torrent. The "private" flag is part of the info block in the torrent file's data. With an infohash mismatch between the two peers no download would happen.
Obviously after a non-compliant party to the transfer has fully downloaded the file(s) it can do whatever it wants with it afterwards… flip any flags and share via DHT, etc.
I recently shared some —more or less— private data to someone else via BitTorrent. We just used DHT for convenience. It took like 15 minutes for other random peers to pop into the transfer. All of those random peers just fetched the meta data. And indeed, a check on btdig confirmed the whole metadata (file names, file sizes, etc.) leaked. So there's a lot of DHT network scanning going on for sure. It was rather fascinating. No actual data was downloaded/leaked at least.
> So there's a lot of DHT network scanning going on for sure.
There is an entire category of free software whose purpose is to create an index of the DHT network. :) The idea is to allow users to find and search for torrents in a completely decentralised manner (i.e. without relying on any centralised trackers or search engines).[1] A good example is bitmagnet[0].
[1] With the added benefit of greater resilience, as centralised "chokepoints" are often the primary and only targets of takedowns.
> So there's a lot of DHT network scanning going on for sure.
How else would btdig (and others) fill their index?
The standard solution is to compress what you're sending with 7zip, with a password.
> No actual data was downloaded/leaked at least.
I've had randos download the data before the intended recipient figured out how to open a port.
I'm the author of https://github.com/anacrolix/torrent and added v2 support a while back.
The short answer is: You just publish as v1 infohash as if nothing is different, but the info contains extra stuff for v2 supporting clients. v1 clients will still work, it's backwards compatible.
So generate your torrent files as hybrid v1/v2, then just do everything else like you always have. Pretend v2 isn't a thing.
v2 clients know how to tap into both swarms and will take advantage of improvements when applicable. It's very well designed.
Thing is, I'm planning on just storing the infohash itself, as a reference (for a knowledge database essentially), not the torrent file, so wanted to use just one of them, and skip the other.
So if I do v1, why do v2? If I do v2, I'd like to skip v1. Ideally I'd do v2 (only) as it seems a heck more future-safe with sha256.
Edit: Actually, I think I see what you're saying. Create hybrid torrent that works with both, share the v1 infohash of that one, v2 clients will automatically take advantage of v2 from that if supported, sounds right?
Love this. I continue to believe Torrenting is the best way to share files, particularly big files, rather than uploading them to some cloud instance with questionable privacy policies…
I used the pipeline
tar | gpg --symmetric | torrent
And now with the added bonus that apparently anything goes, as long as you just say you're downloading training data...
Except when there's no seeds. Usenet seems to have solved this many moons ago.
Is there a self-hostable BitTorrent tracker that can run with Docker? I know there's an option on qBittorrent but I don't want my server to run a full BitTorrent client; just the tracker.
What features are you looking for specifically? If you just want like a really basic index, hosting a JSON file with a object where the key is the infohash and the value is the metadata is a really simple approach :)
No specific feature whatsoever. I basically want anybody that puts <my_bittorrent_tracker_uri> into their torrent tracker list to be able to connect.
That will help me to stop relying on known and popular public trackers[1] in order to share files.
sqtracker https://github.com/tdjsnelling/sqtracker
I've heard of this, but I don't mind anyone with the tracker URI to be able to connect to it. sqtracker is a full-blown system to run a private tracker with user and ratio management, which is overkill in my case.
Others might find the link useful, though!
How about webtorrent's tracker? There is no Docker file but it should be doable https://github.com/webtorrent/bittorrent-tracker
P2P is such an underrated tech
It was never underrated, just too useful. Every time someone invents an actually effective method of person-to-person file transfer, it gets used for piracy and blocked and shunned.
This. I've been working in P2P for 13 years. It's mostly a solved problem. All the new tech and noise are unnecessary. People think improving the tech will cause adoption but it's not the case. P2P remains fringe because it is hard to control and monitor. The friction is social not technical.
It exists on a knife's edge: incredibly useful and powerful to those who are in the know, but obscure enough to avoid heavy handed efforts to obliterate it.
I sometimes wonder how LLMs will impact this. They're much better at surfacing this kind of arcane knowledge than traditional search engines, and that risks increasing accessibility.
If accessibility gets too high, the ISPs could respond against it.
That and it's hard/impossible to monetize, so big corps and VCs are not interested
It's hard to monetize at VC scale, but that doesn't stop small-scale pirates and gray area business to use and make a living out of it.
E.g, in Brazil, Greece and (if you know where to look) even here in Berlin, it is not that difficult to find set top boxes that come with Kodi pre configured with a private tracker and some custom frontend to download movies and shows on demand. In Brazil you buy the box and you pay something like R$200 ($35) per year.
Does a private tracker provide any meaningful security if I'm already encrypting the contents of the torrent?
Remember to disable the dht if you use this.
Unneeded; checking "Private torrent" when creating a torrent means that it won't connect you to the DHT.
For the less technically inclined folks that don't trust this hosted service or can't host their own it's probably simpler to just use a seedbox if you need to access torrented files from multiple locations or you want to share them with others.
This is not what this project is about. You either need a tracker or register to DHT in order to connect to peers. Having a seedbox has nothing to do with that; you'd still either of those for your friends to connect to you and download your files.