Torrent Clients
Seed completed torrents while Plex2Jellyfin imports them.
A torrent stays useful after it finishes. Point your client at a media library root to "import" it, though, and you lose everything Plex2Jellyfin does: filename parsing, canonical naming, placement across drives, the audit log, and the media database. Two clients writing the same library also split one show into several Jellyfin series.
Seeding roots solve this. You give Plex2Jellyfin a directory of completed downloads, it imports from there, and your client keeps seeding the original files.
The contract
Plex2Jellyfin talks to no download client. It watches directories, so any client works if its layout satisfies six rules:
- Incomplete data lives outside every watched root.
- A finished torrent lands in a TV or Movies seeding root.
- The client keeps seeding from that path.
- Plex2Jellyfin never modifies or deletes what it finds there.
- Your client owns ratio limits, seed time, and eventual removal.
- Unclassified downloads land somewhere Plex2Jellyfin does not watch.
One client behaviour breaks this: preallocating the final video filename inside a watched root before the data arrives. Plex2Jellyfin cannot tell a preallocated file from a finished one, so it may import a partial download. Use a separate incomplete directory, or a temporary extension your client renames atomically on completion.
Configuration
[watch]
# Ordinary intake: Plex2Jellyfin moves or copies these, per delete_source.
movies = ["/downloads/complete/movies"]
tv = ["/downloads/complete/tv"]
# Seeding intake: sources always stay in place.
seeding_movies = ["/torrents/complete/movies"]
seeding_tv = ["/torrents/complete/tv"]
[libraries]
movies = ["/media/Movies"]
tv = ["/media/TV Shows"]Ordinary and seeding roots are both intake, so either one pairs with a library. Configure seeding roots alone if every download arrives by torrent.
Roots may not overlap each other or a library path. The daemon rejects such a layout at startup, and all three setup surfaces block it before you save.
Move, copy, hardlink
delete_source in [options] governs ordinary watch roots only:
| Root | delete_source = true | delete_source = false |
|---|---|---|
| Ordinary | Move | Copy |
| Seeding | Hardlink, or copy across filesystems | Hardlink, or copy across filesystems |
A hardlink gives the library a second name for the same data. The file appears in Jellyfin, your client keeps seeding, and the disk holds one copy. When the seeding root and the library sit on different filesystems, Plex2Jellyfin copies instead, so plan for double the space on those titles.
Keep both on one filesystem to get hardlinks. A bind mount does not change this: /torrents and /media hardlink only if they resolve to the same host filesystem.
Permissions
Your torrent client writes the source; Plex2Jellyfin reads it and writes the library. Give both a shared media group, and set the client's umask so group members can read what it creates (002 for group-writable files, 022 when group-read is enough).
Seeding roots need read access, not write. Plex2Jellyfin never touches the source contents.
qBittorrent
Mount staging only. Do not give qBittorrent your media library.
host: /mnt/STORAGE5/.downloads/torrents
container: /downloadsUnder Options → Downloads:
Keep incomplete torrents in: /downloads/incomplete
Default Save Path: /downloads/complete/unclassifiedThen per category (right-click a category → Edit category):
TVSHOWS save path: /downloads/complete/tv
MOVIES save path: /downloads/complete/moviesConfigure seeding_tv and seeding_movies to the category paths. Uncategorised torrents stay in the default path, which nothing watches, so a mislabelled torrent waits for you instead of landing in your library under a guessed name. A manually added torrent imports once it completes and carries one of the two categories.
Transmission
Transmission has no categories. Set an incomplete directory under Preferences → Downloading, then choose the TV or Movies completion root as each torrent's download directory when you add it:
Incomplete directory: /downloads/incomplete
Default directory: /downloads/complete/unclassifiedLeave the default pointing somewhere unwatched. Anything you forget to route stays there.
Deluge
Deluge pairs an incomplete directory with the Label plugin's per-label move-on-complete paths.
Under Preferences → Downloads:
Download to: /downloads/incomplete
Move completed to: /downloads/complete/unclassifiedThen enable the Label plugin and, for each label, set Move completed to:
tv label: /downloads/complete/tv
movies label: /downloads/complete/moviesUnlabelled downloads land in the unwatched default.
Other clients
rTorrent, ruTorrent, and anything else works the same way. Route completed downloads into a media-typed directory, keep partial data elsewhere, and never point the client at a library root. Plex2Jellyfin runs no client-specific code.
Verifying
After the first torrent completes:
plex2jellyfin statusCheck that the file exists in both places and shares an inode with the source:
stat -c '%i %n' /torrents/complete/tv/Show.S01E01.mkv
find /media/'TV Shows' -samefile /torrents/complete/tv/Show.S01E01.mkvMatching inode numbers mean the import hardlinked. Different numbers mean it copied, which is expected across filesystems.