A self-hosted dashboard for your Jellyfin server. See who is watching, keep full play history, dig into statistics and get push notifications on your phone.
Jellydash is a monitoring dashboard for Jellyfin. If you know Tautulli from the Plex world, this is that idea, built for Jellyfin. It runs in Docker next to your server and answers the questions you actually care about:
- Who is watching right now, and is it transcoding?
- What did people watch this week?
- Which shows and movies are the most popular on my server?
- Did someone just request something new in Jellyseerr?
It's supposed to be lightweight, without too much bloat and (hopefully) nice looking!
It also works as a PWA, so you can install it on your phone like a real app. With notifications turned on, your phone buzzes the moment someone hits play. Alerts can go through Telegram, Pushover, Discord or Web Push, whatever you already use. I may work on open-source Android app in the future.
The project is very young and in very active development.
| Now Playing, actively streaming | Now Playing, all quiet |
|---|---|
 |
 |
| History | Jellyseerr requests |
|---|---|
 |
 |
| Trending and Most Watched | Statistics |
|---|---|
 |
 |
Installed as an app on Android, actively streaming or all quiet
More screenshots
-
Now Playing. Live cards for every active stream: artwork, user, quality, progress and the playback method (Direct Play, Remux or Transcode, including the reason why). Live TV channels from tuners like Tunarr show up too, with real program progress and a red on-air badge.
-
History. Every play gets recorded by a background poller, so history is complete even when nobody has the dashboard open. Search it, filter by user or library, export the matching plays to CSV, and enjoy the poster art. Existing Jellyfin Playback Reporting backups can be imported from Settings.
-
Statistics. Watch time trends, top users, device activity, clients, codecs and transcode reasons. There is a Trending strip for what is hot right now, and all-time Most Watched charts for both shows and movies.
-
Libraries. An overview of all your libraries with item counts and type breakdowns. New libraries are picked up automatically.
-
Jellyseerr requests (optional). The latest requests with their current status, plus a push notification when a new request comes in. The page only appears once you connect your Jellyseerr instance.
-
Notifications (optional). "Anna started watching The Office" straight to your phone or desktop, even with the app closed. Delivered through Telegram, Pushover, a Discord webhook, Web Push, or any combination of them.
-
Optional login. Off by default, because on a trusted home network it just gets in the way. One env var turns it on. Recommended if you expose Jellydash to the internet. I recommend using Tailscale for exposing.
-
Modules. Jellydash can load drop-in modules that add whole new pages to the dashboard. See docs/MODULES.md if you want to build your own.
You need Docker with the Compose plugin. Pick the database setup you want, grab two files, and you are ready to go.
The normal setup runs Jellydash and MariaDB together. Choose this if you want the default setup or already use MariaDB.
mkdir jellydash && cd jellydash
curl -LO https://raw.githubusercontent.com/themartz90/jellydash/main/docker-compose.yml
curl -L -o .env https://raw.githubusercontent.com/themartz90/jellydash/main/.env.example
# edit .env, at minimum: JELLYFIN_URL, JELLYFIN_API_TOKEN, DB_PASS
docker compose up -dThe lighter setup runs only Jellydash and keeps its database in ./sqlite-data/jellydash.sqlite. Choose this for a small install where you do not want a separate database container.
mkdir jellydash && cd jellydash
curl -L -o docker-compose.yml https://raw.githubusercontent.com/themartz90/jellydash/main/docker-compose.sqlite.yml
curl -L -o .env https://raw.githubusercontent.com/themartz90/jellydash/main/.env.example
# edit .env, at minimum: JELLYFIN_URL and JELLYFIN_API_TOKEN
docker compose up -dOpen http://your-host:8080 and you are done. Both setups create everything they need automatically, there is nothing to import.
Whichever database you choose, the active setup is saved as docker-compose.yml. Normal commands, aliases and update scripts work the same way for both.
If you want to use your own MariaDB server or mount modules, copy docker-compose.override.example.yml to docker-compose.override.yml and adjust it there.
For setting up notifications, check the section down below.
For both MariaDB and SQLite:
docker compose pull && docker compose up -dIf you installed SQLite using the first v1.2.0 instructions, you may still have a file named docker-compose.sqlite.yml. If there is no docker-compose.yml beside it, make SQLite the default once:
cp docker-compose.sqlite.yml docker-compose.ymlIf docker-compose.yml is your old MariaDB setup, preserve it first:
cp -n docker-compose.yml docker-compose.mariadb.yml
cmp -s docker-compose.yml docker-compose.mariadb.yml && cp docker-compose.sqlite.yml docker-compose.yml && echo "SQLite is now the default Compose setup"These commands only copy the Compose files. They do not change either database. Afterward, the normal update command above manages SQLite.
You do not need to migrate. Existing MariaDB installations continue working exactly as before. Follow this only if you choose to switch to SQLite.
Stop Jellydash first so no plays or settings change during the copy. The migration checks every copied value and leaves the original MariaDB data untouched.
First preserve your current MariaDB Compose file for rollback. This will not overwrite an existing backup. Continue only when it prints MariaDB Compose backup verified:
cp -n docker-compose.yml docker-compose.mariadb.yml
cmp -s docker-compose.yml docker-compose.mariadb.yml && echo "MariaDB Compose backup verified"Then run:
curl -LO https://raw.githubusercontent.com/themartz90/jellydash/main/docker-compose.sqlite.yml
docker compose pull app
docker compose stop app
mkdir -p sqlite-data
docker compose run --rm --no-deps -v "$(pwd)/sqlite-data:/export" app php bin/console.php database:migrate-to-sqlite /export/jellydash.sqlite --confirm-stopped
docker compose down
cp docker-compose.sqlite.yml docker-compose.yml
docker compose up -dFrom then on, the normal docker compose commands manage SQLite. Keep the old MariaDB volume until you have checked the SQLite install and made a backup.
To go back to MariaDB:
docker compose down
cp docker-compose.mariadb.yml docker-compose.yml
docker compose up -dPrefer building the image yourself instead of pulling it from GHCR?
git clone https://github.com/themartz90/jellydash.git
cd jellydash
cp .env.example .env
# edit .env as above
docker compose -f docker-compose.yml -f docker-compose.build.yml up -d --buildUpdating then means git pull and running the same command again. For a source-built SQLite install, replace docker-compose.yml with docker-compose.sqlite.yml.
Jellydash can ping you when someone starts playing and when a new Jellyseerr request comes in. There are four ways to get the alerts, pick whatever you already use. Every configured channel gets every alert, and a channel is on as soon as its values are filled in .env.
Two things that apply to all channels:
- Set
APP_URL=https://your-dashboard.example.comif you want alerts to link back to your dashboard. - You can exclude users from triggering alerts (usually yourself) in the Settings page inside the app.
Test your setup any time, it reports each channel separately:
docker compose exec app php bin/console.php push:test- Message @BotFather, send
/newbotand answer its two questions. It gives you a bot token. - Open a chat with your new bot and send it any message (bots cannot message you first).
- Visit
https://api.telegram.org/bot<YOUR_TOKEN>/getUpdatesin a browser and find"chat":{"id":...}in the response. That number is your chat id! Or you can also use IDbot for your id.
TELEGRAM_BOT_TOKEN=123456789:your-token
TELEGRAM_CHAT_ID=your-chat-idYour User Key is on the pushover.net dashboard. Then create an application there (call it Jellydash) to get an API token.
PUSHOVER_APP_TOKEN=your-app-token
PUSHOVER_USER_KEY=your-user-keyServer Settings > Integrations > Webhooks > New Webhook, pick a channel, copy the webhook URL. No bot needed.
DISCORD_WEBHOOK_URL=https://discord.com/api/webhooks/...The no-third-party option: notifications go straight to your browser or the installed PWA. It needs the dashboard served over HTTPS. Generate a keypair once:
docker compose exec app php bin/console.php push:vapidPaste the two keys into .env, restart, then tap the bell in the app and allow notifications. You get a test notification right away so you know it works.
Set this in .env:
AUTH_ENABLED=true
AUTH_ADMIN_USER=admin
AUTH_ADMIN_PASSWORD=pick-a-strong-oneThe password needs at least 8 characters. The admin user is created automatically on the next start. More users can be added with docker compose exec app php bin/console.php user:add.
Use Export CSV on the History page to choose a search, user, library and time period before downloading. Jellydash shows the exact number of matching plays, and the export is never limited to the page you are viewing. Its versioned format keeps the playback fields Jellydash needs for a native round-trip import. See docs/HISTORY_CSV.md for the v1 format.
To restore that file, open Settings → Import play history and choose Jellydash CSV. Jellydash previews new and already-present plays before asking you to confirm. Imports are transactional, skip duplicates and never trigger playback notifications.
Jellydash only records plays from the moment it starts. If you already used the Playback Reporting plugin, you can import that history.
The plugin backup is a TSV file (no header row): Dashboard → Playback Reporting → Save Backup Data. You can also point at playback_reporting.db in the Jellyfin data folder, or pull live from the plugin API.
Use Import history on the History page, or open the importer directly from Settings. Drop a TSV backup or playback_reporting.db (20 MB max) there. The file type is detected automatically. Jellydash counts the plays first, then asks you to confirm before writing anything. If the plugin is still installed, Import from Jellyfin (API) appears too.
User names are resolved via the Jellyfin /Users API (admin token). Media runtime is looked up from Jellyfin (RunTimeTicks) so the completion bar matches live history; plays are marked finished at 95% of that runtime, same as the poller. If an item no longer exists, runtime stays empty and the play is left unfinished. PlayDuration in the backup is session length, not playback position: a resume that ran 20 minutes shows 20 minutes watched, not 90% of the film. Dates are kept as the plugin recorded them (Jellyfin server local time). Each play is attached to the Jellyfin library that currently owns the item (from its file path); if the item is gone, the type is used as a fallback (Movie → Movies, Episode → TV Shows). Imported plays never trigger notifications. Re-importing skips duplicates, but will fill in a missing runtime and replace a generic library label if Jellyfin is reachable the second time.
- The CPU and RAM numbers in the sidebar come from the host the container runs on.
- iOS only supports Web Push for apps installed to the home screen, and only on iOS 16.4 or newer. That is an Apple rule, not mine 👀. Telegram and Pushover alerts work on any iPhone.
- Brave blocks Web Push by default. It works after enabling "Use Google services for push messaging" in Brave's privacy settings.
- Edge hides Web Push permission prompts behind a small bell icon in the address bar ("quiet notification requests"). If enabling notifications keeps snapping back to off, allow notifications for the site there, and check that Windows itself allows notifications from Edge.
- Trending and Most Watched can exclude libraries you pick (Settings page). Useful for libraries full of temporary stuff.
Jellydash is free and always will be. If you enjoy it and want to support its development, you can buy me a coffee.
The Jellydash mascot is based on a jellyfish icon by Magnific from Flaticon, modified for this project.
Interface icons come from Tabler Icons, used under its MIT license.
The code is licensed under MIT. The mascot original icon is covered by the Flaticon license above, not MIT. Third-party license notices are collected in THIRD_PARTY_NOTICES.md.