Leo
Leo
Docs/home server/media automation/installation/install jellyfin

Last updated Jul 1, 2026

Install & Configure Jellyfin#

This page covers Jellyfin. If you chose Plex instead, see Install & Configure Plex.

With the Jellyfin LXC running and storage mounted, install Jellyfin and connect it to your libraries.

Install Jellyfin on Debian#

Open a shell inside the Jellyfin LXC and run:

apt update && apt upgrade -y
apt install -y curl
 
# Official Jellyfin install script: adds the repository and installs the package
curl https://repo.jellyfin.org/install-debuntu.sh | bash

The script detects your Debian version, adds the Jellyfin apt repository, and installs the package. Jellyfin starts automatically after installation.

Verify it is running:

systemctl status jellyfin

Open the web UI#

From a browser on your LAN:

http://<jellyfin-lxc-ip>:8096

The setup wizard will guide you through initial configuration. Create an admin account; no external account is required.

Add media libraries#

In the Jellyfin setup wizard (or later via Administration → Dashboard → Libraries), add:

Library typePath
Movies/mnt/media/movies
TV Shows/mnt/media/tv

These paths match what was mounted in the Mount Storage step.

Jellyfin will scan on startup. The directories will be empty until Sonarr and Radarr download something.

Automatic library updates#

Jellyfin applies automatic scanning settings to each individual library, not globally. To make Jellyfin detect new files immediately after Sonarr or Radarr imports them, enable real-time monitoring on each library:

  1. Open the Jellyfin Web Client, click the hamburger menu, and select Dashboard
  2. On the left sidebar under the "Server" section, click Libraries
  3. Locate the library you want to update (e.g., Movies or TV Shows), click the three dots (...) on its card, and select Manage Library
  4. Scroll down near the bottom of the settings popup and check Enable real-time monitoring, then click Save

Repeat for both your Movies and TV Shows libraries. Jellyfin will now pick up new files the moment they appear on disk, with no scheduled scan delay.

Note: real-time monitoring relies on inotify and does not work over NFS. If your arr stack and Jellyfin run on separate machines connected by NFS, the filesystem event never reaches Jellyfin and real-time monitoring will not fire.

In that case, use the Sonarr/Radarr built-in Jellyfin connection instead:

  1. In Sonarr or Radarr, go to Settings → Connect and add a Jellyfin connection
  2. Enter the Jellyfin server URL and API key
  3. Enable On Import and On Upgrade

Sonarr/Radarr will notify Jellyfin via API after each import, which works regardless of how Jellyfin accesses the files.

Connect Seerr to Jellyfin#

Seerr supports Jellyfin natively. If you are setting up Seerr for the first time, choose Jellyfin during the setup wizard. If Seerr is already running with Plex, you can add Jellyfin as an additional media server.

  1. In Jellyfin, go to Administration → Dashboard → API Keys and create a key for Seerr
  2. In the Seerr setup wizard, select Jellyfin as your media server
  3. Enter the server URL (http://<jellyfin-lxc-ip>:8096) and the API key
  4. Click Sync Library to confirm the connection

Users authenticate with Jellyfin accounts you create under Administration → Users. No plex.tv account is needed.

GPU passthrough (hardware transcoding)#

Hardware transcoding is free in Jellyfin; no paid tier is required. To enable it, pass through the GPU from the Proxmox host to the Jellyfin LXC.

For Intel iGPU (QSV):

On the Proxmox host, add the following lines to the LXC config file (/etc/pve/lxc/<id>.conf):

lxc.cgroup2.devices.allow: c 226:128 rwm
lxc.mount.entry: /dev/dri/renderD128 dev/dri/renderD128 none bind,optional,create=file
lxc.hook.pre-start: sh -c "chown 100000:100154 /dev/dri/renderD128"

Inside the Jellyfin LXC, add the jellyfin user to the render group:

usermod -aG render jellyfin
apt install -y intel-opencl-icd
systemctl restart jellyfin

For NVIDIA:

NVIDIA passthrough requires more steps than Intel because the NVIDIA userspace libraries inside the container must match the kernel driver version running on the Proxmox host exactly. There is no distribution package that guarantees this match, so the libraries are mounted directly from the host filesystem into the container.

Step 1: Install the NVIDIA driver on the Proxmox host

Follow the standard Debian NVIDIA driver installation on the host. Confirm the driver is loaded:

nvidia-smi

Note the driver version from the output (e.g. 580.142). You will need it in step 3.

Step 2: Configure the LXC

Add the following to /etc/pve/lxc/<id>.conf on the Proxmox host. These lines grant the container access to NVIDIA character devices (major 195), nvidia-uvm (235), DRI (226), framebuffer (29), and input devices (13), then bind-mount the device nodes:

lxc.cgroup2.devices.allow: c 195:* rwm
lxc.cgroup2.devices.allow: c 235:* rwm
lxc.cgroup2.devices.allow: c 226:0 rwm
lxc.cgroup2.devices.allow: c 226:128 rwm
lxc.cgroup2.devices.allow: c 29:0 rwm
lxc.cgroup2.devices.allow: c 13:* rwm
lxc.mount.entry: /dev/dri dev/dri none bind,optional,create=dir
lxc.mount.entry: /dev/nvidia0 dev/nvidia0 none bind,optional,create=file
lxc.mount.entry: /dev/nvidiactl dev/nvidiactl none bind,optional,create=file
lxc.mount.entry: /dev/nvidia-uvm dev/nvidia-uvm none bind,optional,create=file
lxc.mount.entry: /dev/nvidia-uvm-tools dev/nvidia-uvm-tools none bind,optional,create=file
lxc.autodev: 1

Restart the LXC after saving.

Step 3: Create the ldconfig-init.sh entrypoint script

Inside the LXC, create /docker/ldconfig-init.sh. This wrapper runs ldconfig before Jellyfin starts so the container linker picks up the mounted host libraries:

#!/bin/sh
ldconfig
exec /init "$@"

Make it executable:

chmod +x /docker/ldconfig-init.sh

Step 4: Write the Docker Compose file

This approach runs Jellyfin in Docker inside the LXC rather than as a native systemd service. The linuxserver/jellyfin image bundles its own ffmpeg, so the NVIDIA libraries are mounted from the host as read-only volumes. Replace 580.142 throughout with the version from nvidia-smi.

services:
  jellyfin:
    image: lscr.io/linuxserver/jellyfin:latest
    container_name: jellyfin
    privileged: true
    entrypoint: /ldconfig-init.sh
    environment:
      - PUID=0
      - PGID=0
      - TZ=Europe/London
    volumes:
      - /docker/ldconfig-init.sh:/ldconfig-init.sh:ro
      # NVIDIA libraries — version must match nvidia-smi output on host
      - /usr/lib/x86_64-linux-gnu/libnvidia-ptxjitcompiler.so.580.142:/usr/lib/x86_64-linux-gnu/libnvidia-ptxjitcompiler.so.580.142:ro
      - /usr/lib/x86_64-linux-gnu/libnvidia-ptxjitcompiler.so.1:/usr/lib/x86_64-linux-gnu/libnvidia-ptxjitcompiler.so.1:ro
      - /usr/lib/x86_64-linux-gnu/libnvidia-ptxjitcompiler.so:/usr/lib/x86_64-linux-gnu/libnvidia-ptxjitcompiler.so:ro
      - /usr/lib/x86_64-linux-gnu/libnvcuvid.so.580.142:/usr/lib/x86_64-linux-gnu/libnvcuvid.so.580.142:ro
      - /usr/lib/x86_64-linux-gnu/libnvcuvid.so.1:/usr/lib/x86_64-linux-gnu/libnvcuvid.so.1:ro
      - /usr/lib/x86_64-linux-gnu/libnvcuvid.so:/usr/lib/x86_64-linux-gnu/libnvcuvid.so:ro
      - /usr/lib/x86_64-linux-gnu/libcuda.so.580.142:/usr/lib/x86_64-linux-gnu/libcuda.so.580.142:ro
      - /usr/lib/x86_64-linux-gnu/libcuda.so.1:/usr/lib/x86_64-linux-gnu/libcuda.so.1:ro
      - /usr/lib/x86_64-linux-gnu/libcuda.so:/usr/lib/x86_64-linux-gnu/libcuda.so:ro
      - /usr/lib/x86_64-linux-gnu/libnvidia-encode.so.580.142:/usr/lib/x86_64-linux-gnu/libnvidia-encode.so.580.142:ro
      - /usr/lib/x86_64-linux-gnu/libnvidia-encode.so.1:/usr/lib/x86_64-linux-gnu/libnvidia-encode.so.1:ro
      - /usr/lib/x86_64-linux-gnu/libnvidia-encode.so:/usr/lib/x86_64-linux-gnu/libnvidia-encode.so:ro
      # Jellyfin config and media
      - ./library:/config
      - /mnt/jellyfin/jellyfin:/data/tvshows
      - /mnt/jellyfin/movies:/data/movies
      - /mnt/jellyfin/transcodes:/mnt/transcodes:rw
    devices:
      - /dev/dri:/dev/dri
      - /dev/nvidia0:/dev/nvidia0
      - /dev/nvidiactl:/dev/nvidiactl
      - /dev/nvidia-uvm:/dev/nvidia-uvm
      - /dev/nvidia-uvm-tools:/dev/nvidia-uvm-tools
    ports:
      - 8096:8096
      - 8920:8920
      - 7359:7359/udp
      - 1900:1900/udp
    restart: unless-stopped

privileged: true and PUID=0/PGID=0 are required. Unprivileged access to NVIDIA devices from Docker running inside an LXC does not work reliably; the container needs to run as root.

Step 5: Enable NVENC in Jellyfin

After the container starts, go to Administration → Dashboard → Playback → Hardware acceleration and select NVIDIA NVENC. Save and restart the container. Check the transcoding overlay in the Jellyfin player to confirm GPU usage.

Remote access#

Jellyfin does not include a built-in relay. Remote access requires a reverse proxy. If you are already using Cloudflare Tunnels, add a tunnel pointing to http://<jellyfin-lxc-ip>:8096, the setup is the same as for any other service in the stack.

Result#

Jellyfin is installed and connected to your media libraries. No media is available yet; that comes after configuring the download services.

Continue to Configure Download Client (SABnzbd).

Edit on GitHub↗Open an issue↗

← Previous

Install & Configure Plex

Next →

Configure Download Client (SABnzbd)

←All docs