Autopulse + Jellyfin: Instant Library Updates Without Scheduled Scans (2026)

Autopulse + Jellyfin: Instant Library Updates Without Scheduled Scans (2026)

The default Jellyfin workflow for detecting new content is either:

1. Scheduled library scans (every 12 hours by default) - slow, CPU-intensive, and your content sits invisible for hours
2. Radarr/Sonarr Connect notifications - works but requires configuration in each arr app

What if new content appeared in Jellyfin the instant it landed on disk? No scheduled scans. No arr app configuration. Just: file appears > Jellyfin knows about it within 5 seconds.

Autopulse does exactly this.

What Autopulse Does

Autopulse is a lightweight service that:

1. Watches your media folders for new files using inotify (Linux filesystem events)
2. Detects when a file is fully written (waits for write completion, not just creation)
3. Triggers a targeted Jellyfin scan of only the specific folder that changed
4. Debounces rapid changes (multiple files arriving at once trigger one scan, not dozens)

The result: content appears in Jellyfin within 5-15 seconds of the file being fully written to disk.

Why Not Just Use Scheduled Scans?

Autopulse advantages over Radarr/Sonarr Connect:

• Works for any file source (manual copies, Syncthing, Rclone, FTP uploads)
• No configuration needed in each arr app
• Catches files that arr apps miss (manual additions, scripts)
• Single point of configuration

Docker Setup

services:
  autopulse:
    image: ghcr.io/autopulse/autopulse:latest
    container_name: autopulse
    environment:
      - JELLYFIN_URL=http://jellyfin:8096
      - JELLYFIN_API_KEY=your_jellyfin_api_key
      - WATCH_PATHS=/media/movies,/media/tv,/media/music
      - DEBOUNCE_SECONDS=10
      - TZ=Europe/Paris
    volumes:
      - /mnt/media/movies:/media/movies:ro
      - /mnt/media/tv:/media/tv:ro
      - /mnt/media/music:/media/music:ro
    restart: unless-stopped

docker compose up -d

That is it. Autopulse watches the specified paths and triggers Jellyfin scans automatically.

How It Works Internally

1. Filesystem Monitoring (inotify)

Autopulse uses Linux inotify to receive kernel-level notifications when files are created, modified, or moved in the watched directories. This is zero-polling: the kernel tells Autopulse when something changes.

2. Write Completion Detection

When a file is being downloaded (by qBittorrent, Radarr import, or any other process), it is written incrementally. Autopulse waits until the file has not been modified for a configurable period (default: 5 seconds) before considering it "complete."

This prevents triggering a scan on a half-written file.

3. Debouncing

When Radarr imports a movie, it may create the folder, move the file, and write metadata in rapid succession. Without debouncing, this would trigger 3 separate scans.

Autopulse groups changes within a configurable window (default: 10 seconds) and triggers a single scan for the affected path.

4. Targeted Jellyfin Scan

Instead of scanning the entire library, Autopulse calls the Jellyfin API to scan only the specific folder that changed:

POST /Library/Media/Updated
{"Updates": [{"Path": "/media/movies/Dune Part Two (2024)", "UpdateType": "Created"}]}

This is dramatically faster and lighter than a full library scan.

Configuration Options

Ignore patterns

Configure Autopulse to ignore temporary files:

IGNORE_PATTERNS=.partial,.!qB,.tmp,.nfo,Thumbs.db,.DS_Store

This prevents scans triggered by qBittorrent partial files, temporary files, or metadata writes.

Autopulse vs Alternatives

Autopulse is the simplest dedicated solution. If you already have Radarr/Sonarr Connect configured and it works, you may not need Autopulse. But if you add content from multiple sources (manual, scripts, Syncthing), Autopulse catches everything.

Combining with Radarr/Sonarr

You can use both Autopulse AND Radarr/Sonarr Connect notifications:

• Radarr/Sonarr Connect triggers immediately on import (fastest for arr-managed content)
• Autopulse catches everything else (manual additions, other tools)
• Disable scheduled scans entirely (Dashboard > Scheduled Tasks > Scan Media Library > remove triggers)

With both active, you never need scheduled scans again. Every file is detected within seconds regardless of source.

Disabling Scheduled Scans

Once Autopulse is running reliably:

1. Dashboard > Scheduled Tasks
2. Find "Scan Media Library"
3. Remove all triggers (or set to weekly as a safety net)
4. This eliminates the periodic CPU spikes from full library scans

Troubleshooting

Increasing inotify limits

For large libraries with many subdirectories:

echo "fs.inotify.max_user_watches=524288" | sudo tee -a /etc/sysctl.conf
sudo sysctl -p

FAQ

Does Autopulse work with NFS/CIFS mounts? inotify does not work across network filesystems. For NFS/CIFS, Autopulse falls back to periodic polling (configurable interval). Local filesystems get instant inotify events.

Does it work with Emby? Yes. Configure EMBY_URL and EMBY_API_KEY instead of Jellyfin variables. The API call is similar.

How much RAM does Autopulse use? ~20-50 MB. It is extremely lightweight.

Can it trigger other actions besides Jellyfin scans? Some implementations support webhooks, allowing you to trigger any HTTP endpoint when files change.

Should I still keep a weekly scheduled scan as backup? Recommended. Set a weekly scan at 4 AM as a safety net to catch anything Autopulse might miss (edge cases, container restarts).

Content appears instantly. Monitor who watches it from your phone. Download JellyWatch on Google Play - real-time session monitoring for Jellyfin.

On Emby? Download EmbyWatch on Google Play