Like this app? Thanks for giving it a ⭐️

Overview

Automatarr keeps your media library fresh on limited storage by automatically cycling content based on your Radarr/Sonarr import lists.

When content drops off your import lists and isn't in anyone's user pool, it's safely removed — always respecting torrent seeding requirements. Your library stays current without you touching anything.

Each user has a pool of protected content that is immune to automated cleanup. The Discord bot lets users manage their own pools — downloading and removing content without needing access to Radarr/Sonarr or contacting the server owner.

Feature overview:

Loops:

• Library Cleanup - Remove all library content not in Starr App Import Lists while respecting download ratio/time requirements and protecting user pool content.
• Content Search - Search for all wanted missing items across all Starr Apps.
• Queue Cleaner - Monitor and remove all blocked or problematic downloads in the queue.
• Failed Cleanup - Remove all failed downloads from disk.
• Tidy Directories - Remove all unwanted files and directories in the provided paths.
• Permissions Change - Change ownership and permissions of completed downloads.

Bots:

• User Pools - Each user has a pool of content they've downloaded to the server. Users pools are immune to being removed by loops.
• User permission hierarchy - Can assign admins and super users.
• Custom pool size - Each user's limits can be overwritten higher or lower.
• Download - Each user can download x amount of movies with Radarr or series with Sonarr.
• Remove - Each user can remove from their own pool.
• Blocklist - Users can mark a download as unsatisfactory, blocklist it and start a new download.
• Live download status - !list and !waittime show what's really happening: downloading with progress and time left, queued with its place in line, paused, stalled, importing, searching or waiting for release.
• Recommendations - !recommend picks titles on the server you haven't seen, ranked by your taste (e.g. !recommend sci-fi movies). No AI needed.
• Privacy - !private on keeps your Plex watch history and taste out of public replies and popularity counts.

Download clients:

• qBittorrent - Seeding-aware cleanup, so nothing is removed before it has met its seeding requirements.
• SABnzbd (optional) - Discord requests jump to the front of the queue, and real queue positions, progress and time left feed the bot.

Claude AI (optional, bring your own Anthropic API key):

• Chat - Automatarr chats in character when someone @mentions it, replies to it, calls it by name (or a nickname that person gave it), or carries on a conversation after it replies. Everything else is ignored for free.
• Plain-English requests - "Grab me Dune" works from any channel or DM. The download output always lands in the right movie or series channel, folded into one reply.
• Command help - Well-formed ! commands never touch the AI. Malformed or unknown ones are passed to the AI, which runs the corrected command or explains how to type it.
• Knows the server - Finds any title in the library or on TMDB/TVDB with its download status, quality, release dates and who has it, knows your live downloads, and can explain what was removed and why.
• Memory and privacy - Remembers what people like. Users can say "keep my info private", "stop learning about me" or "forget me". Admins can view and delete memories on the Users page.
• Recommendations - From what's on the server that you haven't watched, what's popular this month, or anything else. Proactive ones are rare and event-based: when something new lands that matches someone's taste or when someone comes back after a while.
• Web lookups (optional) - Cast, news and box office the library can't answer, capped at 60 searches a month.
• Budget - A monthly spend cap (default $2.50). If the AI is off, out of credit or over budget, the bot falls back to classic ! commands.

Plex (optional, works with or without the Claude AI):

• Watch activity - Library Cleanup keeps anything someone is watching or watched in the last 14 days, and !blocklist won't delete a file mid-stream.
• Watched in !list - Shows when you last watched each film or which episode you're on (hidden for private users).
• Smarter commands - !stats shows what you've watched and what's unwatched in your pool, !search shows who's watched a title, !popular lists the most watched titles this month, and a full pool suggests watched items to !remove.
• Better AI - Recommendations skip what you've seen, and the AI knows what you're part way through.

Network balancer (optional, needs SABnzbd, qBittorrent and a UniFi gateway):

• One shared budget - Instead of fixed limits that assume both download clients are always flat out, Automatarr shares your connection between them every 10 seconds. Whichever client is busy gets nearly all of it, and an idle one drops to a trickle.
• Household first - Live WAN traffic from your UniFi gateway (UDM, UDM-SE, UCG, UXG) tells Automatarr how much everything else on your network is using. The clients back off within seconds and only take bandwidth back once it has stayed free. A reserve you choose is always kept free on top.
• Read-only UniFi - Use an API key or a local account with the "View Only" role. Automatarr never changes anything on the console.
• Safe if anything stops - The clients' limits never add up to more than your ISP speed minus the reserve, even part way through a change. If Automatarr crashes, a client disappears or the power goes, the network can't be saturated. Losing a download client turns the balancer off and puts both clients on a fixed, even split.
• Takes over the clients' speed settings - Turning it on sets SABnzbd's "Maximum line speed" and qBittorrent's global limits, turns off both apps' speed schedules, and sets qBittorrent's alternative limits very low. Your previous settings are listed on the Network page.

Running Automatarr with Docker Compose:

To run Automatarr using Docker, follow these steps:

1. Make sure you have Docker and Docker Compose installed on any Unix based system

💡 Check with: docker compose version. If your version is below v2.0.0, use docker-compose for commands (with a dash) instead of docker compose.

2. Create a docker-compose.yml File

Create a docker-compose.yml file in your desired directory and add the following content:

services:
  automatarr:
    container_name: automatarr
    image: ghcr.io/maximilian118/automatarr:latest
    restart: unless-stopped
    stop_grace_period: 30s # Time to leave the download clients on safe speed limits when stopping
    ports:
      - "8090:8090" # Frontend
      - "8091:8091" # Backend
    volumes:
      - ./automatarr/database:/app/automatarr_database
      - ./automatarr/logs:/app/automatarr_logs
      - ./automatarr/backups:/app/automatarr_backups
      - /:/host_fs

3. docker compose pull
4. docker compose up -d
5. docker compose logs -f automatarr - All backend information is here
6. Open the web app and start on the Connections page: add Radarr and Sonarr first, then your download clients, then Plex if you use it.

If successful and the application is running, a directory named automatarr will be created alongside the docker-compose.yml file. The automatarr directory contains a database directory where MongoDB stores its local database, as well as a logs directory where all backend logs are stored.

/:/host_fs exposes your machine's entire filesystem to Automatarr. If you're not comfortable with this, that's absolutely fine — simply omit it. However, this means Automatarr will not have access to, and therefore cannot manipulate, content outside of what is achievable through API requests.

Versions and updates

Every push to the main branch is a new version, and the web app shows it in the footer.

Image tag	What it is
:latest	The newest minor or major release (e.g. 0.8.0, 0.9.0). Recommended for most people: fewer, bigger updates.
:0.8 (any X.Y)	The newest patch of that minor version, e.g. 0.8.3.
:0.8.1 (any X.Y.Z)	One exact version, if you want to pin it or roll back.
:edge	Every push to main, as soon as it's built.

To update, run docker compose pull then docker compose up -d. What changed in each version is in CHANGELOG.md and on the Releases page.

Releasing (maintainers): before every push, bump the version in backend/package.json and frontend/package.json (npm version X.Y.Z --no-git-tag-version in each) and add a ## vX.Y.Z section to CHANGELOG.md. The workflow then publishes the images, tags the commit and creates the GitHub Release. It refuses a push whose version was already released.

Connect via Domain (NGINX + SSL)

To access Automatarr via a domain name (e.g. https://automatarr.yourdomain.com), use NGINX or NGINX Proxy Manager to forward traffic to the correct internal ports.

Using NGINX manually:

Replace 192.168.x.x with your server's internal IP, e.g. 192.168.1.100

server {
  listen 443 ssl;
  server_name example.yourdomain.com;

  ssl_certificate     /etc/letsencrypt/live/yourdomain.com/fullchain.pem;
  ssl_certificate_key /etc/letsencrypt/live/yourdomain.com/privkey.pem;

  location / {
    proxy_pass http://192.168.x.x:8090;
    proxy_set_header Host $host;
    proxy_set_header X-Real-IP $remote_addr;
  }

  location /graphql {
    proxy_pass http://192.168.x.x:8091/graphql;
    proxy_set_header Host $host;
    proxy_set_header X-Real-IP $remote_addr;
  }

  location /api {
    proxy_pass http://192.168.x.x:8091/api;
    proxy_set_header Host $host;
    proxy_set_header X-Real-IP $remote_addr;
    proxy_buffering off;
    proxy_read_timeout 1h;
  }
}

Using NGINX Proxy Manager GUI:

Create a Proxy Host for example.yourdomain.com, forwarding to:

Scheme: http
Forward Hostname/IP: 192.168.x.x
Forward Port: 8090

Under the Custom Locations tab, add:

Location: /graphql
Scheme: http
Forward Hostname/IP: 192.168.x.x (Same IP)
Forward Port: 8091

Then click the cog symbol to open the textarea and paste the following:

location /graphql {
  proxy_pass http://192.168.x.x:8091/graphql;
  proxy_set_header Host $host;
  proxy_set_header X-Real-IP $remote_addr;
}

Add a second custom location the same way for the logs page:

Location: /api
Scheme: http
Forward Hostname/IP: 192.168.x.x (Same IP)
Forward Port: 8091

location /api {
  proxy_pass http://192.168.x.x:8091/api;
  proxy_set_header Host $host;
  proxy_set_header X-Real-IP $remote_addr;
  proxy_buffering off;
  proxy_read_timeout 1h;
}

The /graphql and /api locations are required so the frontend can reach the backend through the same domain.

✅ That’s it! You can now access the app securely at https://example.yourdomain.com.

To do:

• Add Discord Bot
• Add Whatsapp Bot
• Add stalled or slow download deletion
• Add graphs for basic data visuals to stats page
• Add Lidarr support
• Add Readarr support
• Add security
• Add webhooks
• Add periodic backups

Legal Disclaimer

Automatarr is a content automation and management tool intended solely for use with legally acquired media.
This software does not host, index, or distribute any content and does not provide or promote access to pirated material.

Automatarr integrates with third-party applications (such as Radarr, Sonarr, and qBittorrent) to help users manage their self-curated media libraries. Any automation involving torrents or downloads is entirely under user control, and it is the user's responsibility to comply with all applicable laws in their country or region.

The developer of Automatarr does not condone or support piracy and is not liable for how others choose to use this software.