Plex MCP Server
Search for movies and manage playlists on your Plex Media Server using the Plex API.
Plex MCP Server
This is a Python-based MCP server that integrates with the Plex Media Server API to search for movies and manage playlists. It uses the PlexAPI library for seamless interaction with your Plex server.
Screenshots
Here are some examples of how the Plex MCP server works:
1. Find Movies in Plex Library by Director
Search for movies in your Plex library by specifying a director's name. For example, searching for "Alfred Hitchcock" returns a list of his movies in your library.

2. Find Missing Movies for a Director
Identify movies by a specific director that are missing from your Plex library. This helps you discover gaps in your collection.

3. Create a Playlist in Your Plex Library
Create a new playlist in your Plex library using the movies found in a search. This allows you to organize your library efficiently.

Setup
Prerequisites
- Python 3.8 or higher
uvpackage manager- A Plex Media Server with API access
Installation
Installing via Smithery
To install Plex Media Server Integration for Claude Desktop automatically via Smithery:
npx -y @smithery/cli install @djbriane/plex-mcp --client claude
Installing Manually
-
Clone this repository:
git clone <repository-url> cd plex-mcp -
Install dependencies with
uv:uv venv source .venv/bin/activate uv sync -
Configure environment variables for your Plex server:
PLEX_TOKEN: Your Plex authentication tokenPLEX_SERVER_URL: Your Plex server URL (e.g., http://192.168.1.100:32400)
Finding Your Plex Token
You can find your Plex token in this way:
- Sign in to Plex Web App
- Open Developer Tools
- In Console tab, paste and run:
window.localStorage.getItem('myPlexAccessToken')
Usage with Claude
Add the following configuration to your Claude app:
{
"mcpServers": {
"plex": {
"command": "uv",
"args": [
"--directory",
"FULL_PATH_TO_PROJECT",
"run",
"src/plex_mcp/plex_mcp.py"
],
"env": {
"PLEX_TOKEN": "YOUR_PLEX_TOKEN",
"PLEX_SERVER_URL": "YOUR_PLEX_SERVER_URL"
}
}
}
}
Available Commands
The Plex MCP server exposes these commands:
| Command | Description | OpenAPI Reference |
|---|---|---|
search_movies | Search for movies in your library by various filters (e.g., title, director, genre) with support for a limit parameter to control the number of results. | /library/sections/{sectionKey}/search |
get_movie_details | Get detailed information about a specific movie. | /library/metadata/{ratingKey} |
get_movie_genres | Get the genres for a specific movie. | /library/sections/{sectionKey}/genre |
list_playlists | List all playlists on your Plex server. | /playlists |
get_playlist_items | Get the items in a specific playlist. | /playlists/{playlistID}/items |
create_playlist | Create a new playlist with specified movies. | /playlists |
delete_playlist | Delete a playlist from your Plex server. | /playlists/{playlistID} |
add_to_playlist | Add a movie to an existing playlist. | /playlists/{playlistID}/items |
recent_movies | Get recently added movies from your library. | /library/recentlyAdded |
Running Tests
This project includes both unit tests and integration tests. Use the following instructions to run each type of test:
Unit Tests
Unit tests use dummy data to verify the functionality of each module without requiring a live Plex server.
To run all unit tests:
uv run pytest
Integration Tests
Integration tests run against a live Plex server using environment variables defined in a .env file. First, create a .env file in your project root with your Plex configuration:
PLEX_SERVER_URL=https://your-plex-server-url:32400
PLEX_TOKEN=yourPlexTokenHere
Integration tests are marked with the integration marker. To run only the integration tests:
uv run pytest -m integration
If you are experiencing connection issues to your Plex server try running the integration tests to help troubleshoot.
Code Style and Conventions
-
Module Structure:
Use clear section headers for imports, logging setup, utility functions, class definitions, global helpers, tool methods, and main execution (guarded byif __name__ == "__main__":). -
Naming:
Use CamelCase for classes and lower_snake_case for functions, variables, and fixtures. In tests, list built-in fixtures (e.g.monkeypatch) before custom ones. -
Documentation & Comments:
Include a concise docstring for every module, class, and function, with in-line comments for complex logic. -
Error Handling & Logging:
Use Python’sloggingmodule with consistent error messages (prefix “ERROR:”) and explicit exception handling. -
Asynchronous Patterns:
Define I/O-bound functions as async and useasyncio.to_thread()to handle blocking operations.
Servidores relacionados
reachy-mini-mcp
Control the Reachy Mini robot (or simulator) from Claude, ChatGPT, or any MCP-compatible client.
Lichess MCP
Interact with the Lichess chess platform using natural language.
SciPilot
Natural language interface for scientific command-line tools via MCP
Umami MCP server
MCP server exposing Umami analytics (Cloud + self-hosted)
mcp-dice
Rolls dice using standard notation (e.g., 1d20) and returns individual rolls and their sum.
MCP-Ambari-API
Automate Apache Ambari operations with AI/LLM: Natural language commands for Hadoop cluster management, service control, configuration monitoring, and real-time status tracking via Model Context Protocol (MCP) tools.
StonkWatch
Real-time ASX market intelligence for AI agents — announcements, AI summaries, sentiment, social intelligence, stock prices, and franking credit calculator across 2,200+ Australian-listed companies.
Amazon Seller MCP Server
Connect Amazon Seller Central to Claude or ChatGPT via Two Minute Reports MCP and get accurate insights on orders, sales, inventory, and revenue performance.
PaKi Curator — Visual Medicine Art Catalog
MCP server for César Yagüe's Visual Medicine — 301 contemplative moving art works (seamless loops up to 16K), 109 vibrational capsules, 6 tools including spatial recommendations and direct artist inquiry.
Baozi Bet Prediction Markets
Baozi bet MCP server to allow agents create their own prediction markets