Plex Movie Recommendations Script Triggered from Tautulli within Docker

Table of Contents

• Overview
• Features
• Requirements
• Installation-Setup
  • 1. Prerequisites
  • 2. Prepare Your config.yaml
  • 3. Build the Docker Image
  • 4. Redeploy Tautulli with the Custom Image
  • 5. Set Up Tautulli Automation

---

Overview

This script automates the process of:

1. Generating movie recommendations via OpenAI (with TMDb fallback) .
2. Checking Plex for existing recommendations.
3. Adding missing movies to Radarr or Overseerr (user-configurable).
4. Maintaining a dynamic Plex collection ("Inspired by your Immaculate Taste")
5. Using a points system to keep recommendations fresh and relevant
6. Clean up duplicates in Plex and Unmonitored in Radarr

---

Features

• GPT Recommendations:
  Generates up to 50 (configureable) movie suggestions based on a “seed” title, including sequels, same-director picks, similar genres, hidden gems, etc.

• Plex Integration:

  • Searches Plex for each recommended title.
  • If found, the script will later include it in (or keep it in) a dedicated collection.
  • Remove one file from movies identified as duplicates

• Radarr Automation:

  • If a recommended title is not in Plex, the script adds it to Radarr with a configurable root folder and tag, so Radarr can download it.
  • Unmonitor the movies already existing in Plex

• Overseerr Support:

  • Modified version of Overseerr needed GitHub_Overseerr_Modified otherwise all the recommendations sent to Overseerr will automatically get approved. Thats just how Overseerr is designed, Requests made via API are directly approved. GitHub Overseerr_Issue_3926
  • Submit recommendations for manual approval
  • Configurable root folders and quality profiles

• Points System:

  • New recommendations get +10 points.
  • Each run, existing items lose 1 point.
  • Movies must have ≥5 points or a TMDb rating >8 to remain in the collection.
  • Points are persisted across runs in a JSON file.

• TMDb Caching & Ratings:

  • Caches TMDb IDs and ratings locally to reduce external API calls.
  • Uses each movie’s rating (>8) to help decide if it stays in the collection despite low points.

• YAML Configuration:

  • No hardcoding credentials.
  • All API keys, file paths, and server URLs are loaded from config.yaml.

---

Requirements

1. Core Services:
   Docker
   Plex
   Tautulli
   
2. APIs:
   TMDb API Key
   OpenAI API Key (Optional)
   
3. Download Management: (At Least One)
   Radarr OR
   Overseerr

---

Installation-Setup

This project includes a Dockerfile based on lscr.io/linuxserver/tautulli:latest. It installs the necessary Python libraries (e.g., openai, arrapi, PyYAML) and allows you to run the recommendation script inside the Tautulli container.

1. Prerequisites

• Docker, Plex, Tautulli, and Radarr must already be installed and working.
• You’ll need valid credentials for each service (tokens, API keys, etc.).

2. Prepare Your config.yaml

1. Create or edit config.yaml with your real Plex, OpenAI, Radarr, Overseerr and TMDb credentials.
2. Make sure config.yaml is placed alongside your Docker build or mounted at runtime (see below).

3. Build the Docker Image

From the directory "Tautulli_Curated_Plex_Collection/custom-tautulli/" that contains the Dockerfile, run the command in terminal:

docker build -t tautulli_recommendations .

4. Redeploy Tautulli with the Custom Image

If you removed the old container, you’ll now create a new one. Follow the steps based on whether you're using Portainer or Docker Compose.

• Option 1: Using Portainer

1. In Portainer, choose Add Container (if you removed the old one) or Duplicate/Edit (if you’re editing the existing container).
2. In the Image field, enter the name of your custom image:

tautulli_recommendations:latest

3. Proceed to configure ports, volumes, and environment variables as needed, then Deploy the container.

---

• Option 2: Using Docker Compose (No Portainer)
• If you are managing Tautulli with Docker Compose, follow these steps:

1. Stop the existing Tautulli container:

docker compose stop tautulli

2. Build the new custom image:

docker build -t tautulli_recommendations .

3. Start Tautulli with Docker Compose:

docker compose up -d tautulli

Once complete, Tautulli will be running with your custom-built image and ready to use.

5. Set Up Tautulli Automation

To have Tautulli automatically call your script whenever someone finishes watching a movie (or meets another trigger):

1. Open Tautulli → Settings → Notification Agents.
2. Click Add a new notification agent and choose Script.
3. Script Folder: Browse to the folder where you stored the recommendation script (e.g., /app if that’s where tautulli_watched_movies.py resides).
4. Script File: Select the script file you want Tautulli to run (e.g., tautulli_watched_movies.py).
5. Description: Provide a friendly name (e.g., “Movie Recommendation Script”).
6. Trigger: Choose Watched (so the script runs when a user finishes watching a movie).
7. Arguments: Under Watched arguments, pass "{title}" (including quotes), so the script receives the movie’s title. Example:

    python tautulli_watched_movies.py "{title}"

8. Test Notification: Click Test → select your script → provide "Inception (2010)" as the argument. This simulates a watch event for “Inception (2010).”
9. Verify: Check Tautulli’s logs to see if the script ran successfully and see any output (e.g., “Found in Plex already,” “Added to Radarr,” etc.).

---

Now Whenever Tautulli detects that a user has finished watching a movie, it will trigger your script with the movie’s title. With each run, your collection becomes more finely curated.

Tip: Add the collection to your Home screen and position it at the very top—right beneath the Continue Watching list.

Enjoy using this script! I hope it enhances your movie selection. If you Encounter any issues or have ideas for enhancements? Feel free to open an issue or submit a pull request.

---

License

This project is provided “as is” without warranty of any kind. You are free to use, modify, and distribute this code as per the MIT License.

---

View of Collection in Plex App:

Logs from Tautulli after a test run:

Overseerr Approval request

Plex Duplicate Cleanup

Radarr Unmonitor Existing Plex Movies