Match torrents and data, remove torrents based on data, cleanup your disk for unseeded files. Autotorrent2 does everything you currently miss in your flow.

• Torrent clients: rtorrent, Deluge, Transmission and qBittorrent
• OS: Any, but only tested on linux
• Python: 3.7+ are the only tested versions, might work with lower 3.x.

Install Autotorrent2

python3 -m venv ~/.autotorrent # Create virtual environment where we install autotorrent2
~/.autotorrent/bin/pip install autotorrent2 # Actually install autotorrent2

# Optional, add at2 to your commandline
echo "alias at2=~/.autotorrent/bin/at2" >> ~/.bashrc
source ~/.bashrc

The first time Autotorrent2 is run without a known config file, it will create a base config file.

at2 check-config

The default location is ~/.config/autotorrent/config.toml - edit it to match your setup. See the example configuration file for setting description found here.

Test the connections and see if it can connect to all your configured clients.

at2 test-connection

Now you are ready to actually use it, check out the Usage page for fun things to do and CLI page for featurelist

It can be fun to run scripts automatically and see cross-seeding just happen. Autotorrent2 is not really designed for multiple instances at once so it is recommenced to use a lock to prevent this.

Instead of just at2 then use e.g. flock ~/.autotorrent.lock -c 'at2' which prevents multiple instances of Autototrrent2 at once.

If you use Autotorrent2 in a docker container or with a torrent client running in docker then the recommendation is to align the paths. If your torrent data is located in /mnt/data outside docker then you should map it similarly inside the docker container and you will save yourself from a lot of headaches.

Personally I recommend mounting your data source as read-only because bittorrent clients are write-happy that might corrupt your data.

Assortment of stuff that is planned.

• When Autotorrent2 is working on a task, e.g. copying a file to cache, then it might look like as it is stalled. An indicator should be added.
• Client normalization indexing, e.g. index based on how transmission and qbittorrent handles problematic filenames
• Torrent discovery for a torrent site.

Assortment of stuff I am not sure I can do much about.

• Transmission (3.x) does not parse all emojis correctly and will return the wrong filename

MIT