jellytui

A very opinionated Shows-only TUI for Jellyfin.

Features

• 3-pane layout showing Shows > Seasons > Episodes, you can navigate this with your arrow keys
• each episode displays its runtime in mins:secs
• ability to spawn an mpv player for watching the selected episode by pressing s. Upon closing mpv, the point where you left off is saved on your jellyfin server so you can resume Every 15 seconds your watch progress is updated on your jellyfin server.
• mark episode as watched/unwatched manually with space (if you've watched most of it, jellyfin should mark it as watched automatically)
• force refresh a show's cache with r

The Caching Mechanism

Note

The cache is stored in

• linux: ~/.local/share/jellytui/state.json
• windows: ~/AppData/Roaming/jellytui/jellytui/data

In order to not bomb my server with requests, I tried minimizing network interactions.

• when you launch the TUI, all the data you see immediately displayed is pulled from the cache (in subsequent runs of course).
• when you launch the TUI, all the show names are fetched (the left pane)
  • since the first show is selected by default, this is also fetched here
• as you scroll down each show, that specific show is fetched. This is done only once per run to limit the amount of requests. The idea here being that you probably won't be updating the shows that much while watching them but if you are you can force reload the specific show with r
• when you exit, the data will be saved in the cache file

This makes it so everything on screen is by default only fetched once.

Environment Variables

You can see what's needed in .example.env. Either create a .env file or pass these as arguments.

• JELLYFIN_URL: Should look something like this http://192.168.0.63:8096
• JELLYFIN_API_KEY: Open any episode and click More > Copy stream URL. At the end of the link you should see ?api_key=..., paste that here.
• JELLYFIN_USER_ID: Go to your profile and paste your ?userId=...

State of the Project

I mean it works obviously but as I said earlier, this was made specifically to suit me and not to be the best looking or most functional or most configurable for any user. I likely won't take feature requests, maybe bug reports (unless its something important), I encourage you to take the code and make it your own if you want to :)

A Disclaimer on AI Use

I recently realized I had access to Gemini Pro via my university and wanted to see how relying on an LLM for programming is. As a result maybe 90% of the code here is AI generated. Can't say I was a big fan and likely won't lean on LLMs to this degree again in the future. Might write a short blog post about this at some point.