- Home
- Skills
- Agents & Orchestration
- Jellyfin MCP – Control your media server with AI
Works with the AI Agents you already use
Jellyfin MCP – Control your media server with AI
Turn your AI agent into a smart remote and librarian for your **own Jellyfin media server**.
Free
Jellyfin MCP – Control your media server with AI
Example session with this skill installed
Find an unwatched Murnau film in my library and play it on the living room TV.
- Read your context and instructions
- Compiled the jellyfin mcp
Illustrative example on a sample library.
Summary: I looked up Murnau in your library, kept only the films you haven't watched, and started one on your living room TV.
1. Unwatched Murnau films in your library
Called jellyfin_get_person with name: "Murnau" and filtered the returned items on played: false:
| Title | Year | Runtime | Status |
|---|---|---|---|
| Sunrise: A Song of Two Humans | 1927 | 1h 34m | Unwatched |
| Faust | 1926 | 1h 56m | Unwatched |
| Nosferatu | 1922 | 1h 34m | 20 min in (resume available) |
You asked for one film, so I picked Sunrise (fully unwatched). Faust is next in line.
2. Target device
Called jellyfin_list_sessions:
| Device | Client | Now playing |
|---|---|---|
| Living Room TV | Jellyfin for Android TV | Idle |
| Pixel 8 | Jellyfin Mobile | Idle |
Exactly one device matches "living room", so no need to ask.
3. Playback
Called jellyfin_play with the item id from step 1, the Living Room TV session id and mode: now.
Sunrise: A Song of Two Humans (1927) is now playing on Living Room TV.
Next steps
- Say "pause" or "skip ahead 10 minutes" and I'll use
jellyfin_playback_control. - Want Faust queued after it? I can call
jellyfin_playwithmode: next. - I can also add the remaining Murnau films to a "Silent classics" playlist.
Connects securely to your tools. The creator never sees your data.
About this skill
Turn your AI agent into a smart remote and librarian for your own Jellyfin media server.
This skill teaches your agent how to use the jellyfin_* tools of the free, open-source Jellyfin MCP Server plugin correctly: it finds titles, recommends only what is actually in your library, starts playback on the right TV or device, manages playlists, favorites and watched state, and runs admin tasks.
What it does well
- Never guesses ids. Every item, person, playlist and session id comes from the tools, so no "item not found" dead ends.
- Recommends from your library. Uses your next up, continue watching, unplayed items and watch history, never titles you don't own.
- Picks the right device. Lists active sessions and asks which device when several match before calling play.
- Handles paging honestly. Respects
totalCountinstead of presenting the first page as the whole library. - Asks before irreversible actions. Deleting playlists, clearing watched progress, running maintenance tasks or replacing metadata always need your confirmation.
Quick reference
| You want | The agent calls |
|---|---|
| Find a title | jellyfin_search |
| Details, cast, overview | jellyfin_get_item |
| Everything with an actor/director | jellyfin_get_person |
| "What should I watch?" | jellyfin_next_up, jellyfin_continue_watching, jellyfin_recently_added |
| Play on a device | jellyfin_list_sessions + jellyfin_play |
| Pause, seek, skip | jellyfin_playback_control |
| Playlists, favorites, watched | jellyfin_manage_playlist, jellyfin_set_favorite, jellyfin_set_played |
| Admin | jellyfin_run_task, jellyfin_scan_libraries, jellyfin_refresh_item |
Requirements
- A Jellyfin server with the Jellyfin MCP Server plugin installed (free, GPL-3.0): https://github.com/AeLui/jellyfin-plugin-mcp
- A per-user token created in Dashboard > Plugins > MCP Server, connected to your agent as an MCP server. Every call runs as that Jellyfin user and only within the token's scopes (
library,playback,manage,admin).
The skill itself runs no code; all actions go through your own Jellyfin MCP tools.
Requires the Jellyfin MCP Server plugin on your Jellyfin server and a token connected to your agent. Only the tools of the token's scopes are available, and everything runs with that Jellyfin user's permissions. Playback needs an open, controllable Jellyfin client on the target device. Plugin updates cannot be triggered through admin tasks (refused by design).
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 9 days ago
- Free with an Agensi account
Requires
Permissions
Allowed Hosts
The skill itself runs no code; all actions go through the user's own Jellyfin MCP tools.