Skip to content


What is it?

Autoscan replaces the default Plex, Emby, and Jellyfin behaviour for picking up file changes on the file system. Autoscan integrates with Sonarr, Radarr, Lidarr and Google Shared Drives to fetch changes in near real-time without relying on the file system.

Autoscan is a rewrite of the original Plex Autoscan written in the Go language. In addition, this rewrite introduces a more modular approach and should be easy to extend in the future.

Project home Docs Github Docker


The Plex API is known to have trouble when scanning items into empty libraries. You should add at least one item to each Plex library and perform a manual scan as a first step. If you don't do this, things may not get scanned into Plex in response to autoscan's requests.

The Saltbox Autoscan role will attempt to partially configure your autoscan config file located at /opt/autoscan/config.yml. You should refer to the documentation and adjust this file as suits your own needs. The config generated is very minimal. a-train is now replacing the bernard trigger.

The generated config file will look something like this:

# <- processor ->

# Override the minimum age before a scan request is sent to the target (Default 10m):
minimum-age: 10m

# Override the delay between processed scans (Default 5s):
scan-delay: 5s

# Set anchor files for remote storage. If these are missing no scans will be sent to the target to avoid files being trashed when a mount fails
  - /mnt/unionfs/mounted.bin

# <- triggers ->

# Optionally, protect your webhooks with authentication

# Port for Autoscan webhooks to listen on
port: 3030

      priority: 5
      rewrite: # Global rewrites
        - from: ^/Media/
          to: /mnt/unionfs/Media/

    - priority: 0

      # Filter with regular expressions
        - ^/mnt/unionfs/Media/
        - '\.(srt|pdf)$'

      # rewrite inotify path to unified filesystem
        - from: ^/mnt/local/Media/
          to: /mnt/unionfs/Media/

      # Local filesystem paths to monitor
        - path: /mnt/local/Media

    - name: sonarr # /triggers/sonarr
      priority: 2

    - name: radarr # /triggers/radarr
      priority: 2

    - name: lidarr # /triggers/lidarr
      priority: 1

# <- targets ->

    - url: https://plex.DOMAIN.TLD # plex
      token: YOUR_PLEX_TOKEN

Then edit the anchors section:

  - /mnt/unionfs/mounted.bin

To reflect your own configuration.

For example, if you went through the saltbox rclone setup process, you'll need to enter something like this:

  - /mnt/unionfs/bvoiwepopz-movies_mounted.bin
  - /mnt/unionfs/bvoiwepopz-tv_mounted.bin
  - /mnt/unionfs/bvoiwepopz-music_mounted.bin
  - /mnt/unionfs/bvoiwepopz-anime_mounted.bin

You should enter the entire list of bin files that were created by the automated script here.

ls /mnt/unionfs/*.bin will give you the list of files you should enter here.

Everything else should be ready to go for standard usage.

What are those mount files?
Autoscan uses these to determine if your cloud storage is mounted and visible; if autoscan can't see these files, no scans will be sent to Plex since doing so would empty your library as Plex removed all the files it can no longer see [assuming that "empty trash on scan" is enabled].
There's nothing special about the contents of these files; autoscan just needs to see that they exist. Typically they are empty.
If you went through the saltbox rclone setup, these files got created for you.
Do I really need to include all seven or eight or however many?
Strictly speaking, no, not with the way saltbox sets up the mounts. All those shared drives are part of a union remote, and the union remote is mounted, so there's really no possibility that some of those files would be present but not others. Any one of them is probably sufficient.
However, there's no reason *not* to include them all as you can grab the list with a single command and a copy-paste. You save a few keystrokes by not including all of them [you don't have to copy-paste ` - ` in front of those few lines], but in thinking about it at all you've spent the same amount of time. Reading this question and answer have taken more time than it would have taken to include all of them as a belt-and-suspenders measure.

You will set up the webhooks for radarr/sonarr/lidarr as part of their setup, so they aren't discussed here

Manual Scan URL

The manual scan URL will be https://autoscan.YOUR_DOMAIN/triggers/manual. Usage is described in the autoscan docs linked below.


Autoscan can monitor Google Drive changes via a trigger called "Bernard". The code behind Bernard can sometimes get out of sync with the state of Google Drive and miss things, so now we are using A-Train.

IMPORTANT: You only need to set this up if you are planning to add media to Google Drive directly, outside the usual Radarr/Sonarr channels, or if you are monitoring a Shared Drive where new media appears outside those channels. If you are not planning to do that, you can skip this portion of the setup.

"A-Train" is a rewrite of the Bernard concepts, and is currently available as a second docker image as part of Sandbox. It will likely be integrated into autoscan at some point in the future.


A-Train supports only unencrypted Google Shared Drives authenticated via Service Accounts. It does not support encrypted drives, My Drive, or authentication via Client ID/Secret or other means.

Enter the names of the remotes you want to monitor in the sandbox settings.yml. The Remotes can be either drive remotes or union remotes. You may use rclone listremotes to get your drive remotes.


  remotes: ["bvoiwepopz-Movies", "bvoiwepopz-TV"]


  remotes: ["google"]

Run the a-train tag to create the container:

sb install sandbox-a_train

Copy one of your service account files from its current location to /opt/a-train/account.json. Remember to rename your service account file to "account.json".


cp /opt/sa/all/160.json /opt/a-train/account.json

Run the autoscan tag to rebuild the container:

sb install autoscan

Run the a-train tag to rebuild the container:

sb install sandbox-a_train


If for some reason you still wanted to use Bernard, it would look like this:

    - account: /config/sa.json # Path inside the container where your SA is located
      cron: "*/5 * * * *" # every five minutes (the "" are important)
      priority: 0
        - id: drive_id #Friendly title
      # Rewrite gdrive to the local filesystem
        - from: ^/Media/
          to: /mnt/unionfs/Media/
      # Filter with regular expressions
        - ^/mnt/unionfs/Media/
        - '\.srt$'

Further documentation:


Are you setting Saltbox up for the first time? Continue to Sonarr.