Skip to content

Deployment

Deployment Options

jellyfin-mcp supports local stdio, a loopback-only development listener, a least-privilege stdio container, and a remote authenticated HTTPS boundary. Provider endpoint, credential, selector, identity, and trust material are supplied at runtime through AgentConfig; none is stored in this repository.

Installed stdio process

{
  "mcpServers": {
    "jellyfin": {
      "command": "jellyfin-mcp",
      "args": [],
      "env": {"MCP_TOOL_MODE": "intent"}
    }
  }
}

Loopback development listener

jellyfin-mcp --transport streamable-http --host 127.0.0.1 --port 8000

Do not expose this listener beyond loopback. Network deployments require direct TLS or an explicitly trusted TLS-terminating ingress, configured authentication, exact MCP_ALLOWED_HOSTS, and an exact trusted-proxy CIDR policy.

Least-privilege local container

docker run -i --rm \
  --read-only \
  --cap-drop=ALL \
  --security-opt=no-new-privileges \
  --pids-limit=256 \
  --tmpfs /tmp:rw,noexec,nosuid,nodev,size=64m \
  -e TRANSPORT=stdio \
  registry.example.invalid/jellyfin-mcp@sha256:<digest> jellyfin-mcp

The operator projects the selected AgentConfig profile into the process at runtime; the image remains immutable and contains no environment connection profile.

Remote authenticated HTTPS endpoint

{
  "mcpServers": {
    "jellyfin": {"url": "https://service.example.invalid/mcp"}
  }
}

Store the real remote URL, outbound identity reference, and TLS-profile reference in AgentConfig, not in MCP client JSON or documentation.

This page covers running jellyfin-mcp as a long-lived server: the transports, a Docker Compose stack, the integrated A2A agent, putting it behind a Caddy reverse proxy, and giving it a DNS name with Technitium. To provision the Jellyfin media server it connects to, see Backing Platform.

jellyfin-mcp ships two servers: an MCP server (console script jellyfin-mcp) and an A2A graph agent (console script jellyfin-agent). The MCP server is a typed, deterministic tool surface; the agent wraps it with a Pydantic-AI graph, an optional Web UI, and OpenTelemetry tracing.

Run the MCP server

The transport is selected with --transport (or the TRANSPORT env var):

jellyfin-mcp
For IDE / desktop MCP clients that launch the server as a subprocess.

jellyfin-mcp --transport streamable-http --host 0.0.0.0 --port 8000
A network server with a /health endpoint and /mcp route.

jellyfin-mcp --transport sse --host 0.0.0.0 --port 8000

Health check (HTTP transports):

curl -s http://localhost:8000/health        # {"status":"OK"}

Configuration (environment)

jellyfin-mcp is configured entirely from the environment. The required connection set:

Var Default Meaning
JELLYFIN_URL http://localhost:8096 Jellyfin server base URL
JELLYFIN_API_KEY (unset) API key / access token (preferred auth)
JELLYFIN_USERNAME admin Username (credential login, if no key)
JELLYFIN_PASSWORD (unset) Password (credential login, if no key)
JELLYFIN_TLS_PROFILE (unset) Named runtime TLS profile
JELLYFIN_TLS_PROFILE_REF (unset) Secret reference for private PKI, mTLS, or proxy policy
AUTH_TYPE apiKey Authorization flow: apiKey, credentials, or delegated
CONDENSED_JELLYFIN_TOOL True Register the condensed Jellyfin tool set

Provide either JELLYFIN_API_KEY or JELLYFIN_USERNAME + JELLYFIN_PASSWORD; when neither is present the server remains inactive rather than failing at import time. Optional OIDC token delegation (ENABLE_DELEGATION, OIDC_TOKEN_ENDPOINT, OIDC_CLIENT_ID, OIDC_CLIENT_SECRET, JELLYFIN_AUDIENCE, DELEGATED_SCOPES) and telemetry (ENABLE_OTEL, OTEL_EXPORTER_OTLP_*) settings are documented in .env.example. Copy it to .env and populate only what you use. Plus HOST / PORT / TRANSPORT for HTTP transports.

Docker Compose

The repo ships docker/mcp.compose.yml. It reads a sibling .env and publishes the HTTP server on :8000:

services:
  jellyfin-mcp-mcp:
    image: example/jellyfin-mcp@sha256:<digest>
    container_name: jellyfin-mcp-mcp
    hostname: jellyfin-mcp-mcp
    restart: always
    env_file:
      - ../.env
    environment:
      - PYTHONUNBUFFERED=1
      - HOST=0.0.0.0
      - PORT=8000
      - TRANSPORT=streamable-http
    ports:
      - "8000:8000"
    healthcheck:
      test: ["CMD", "python3", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:8000/health')"]
      interval: 30s
      timeout: 10s
      retries: 3

cp .env.example .env          # then edit JELLYFIN_* values
docker compose -f docker/mcp.compose.yml up -d
docker compose -f docker/mcp.compose.yml logs -f

Agent server

The integrated A2A graph agent is published as the jellyfin-agent console script. It connects to the MCP server over HTTP (MCP_URL), exposes a Web UI / Agent Control Protocol surface, and is driven by a configurable LLM provider.

export JELLYFIN_URL=http://your-jellyfin:8096
export JELLYFIN_API_KEY=your_api_key
jellyfin-agent --provider openai --model-id gpt-4o --host 0.0.0.0 --port 9056

The repo ships docker/agent.compose.yml, which runs the MCP server and the agent together. The agent listens on :9056 and is wired to the MCP server by container name:

services:
  jellyfin-mcp-mcp:
    image: example/jellyfin-mcp@sha256:<digest>
    container_name: jellyfin-mcp-mcp
    hostname: jellyfin-mcp-mcp
    restart: always
    env_file:
      - ../.env
    environment:
      - PYTHONUNBUFFERED=1
      - HOST=0.0.0.0
      - PORT=8000
      - TRANSPORT=streamable-http
    ports:
      - "8000:8000"

  jellyfin-mcp-agent:
    image: example/jellyfin-mcp@sha256:<digest>
    container_name: jellyfin-mcp-agent
    hostname: jellyfin-mcp-agent
    restart: always
    depends_on:
      - jellyfin-mcp-mcp
    env_file:
      - ../.env
    command: [ "jellyfin-agent" ]
    environment:
      - PYTHONUNBUFFERED=1
      - HOST=0.0.0.0
      - PORT=9056
      - MCP_URL=http://jellyfin-mcp-mcp:8000/mcp
      - PROVIDER=${PROVIDER:-openai}
      - MODEL_ID=${MODEL_ID:-gpt-4o}
      - ENABLE_WEB_UI=True
      - ENABLE_OTEL=True
    ports:
      - "9056:9056"

docker compose -f docker/agent.compose.yml up -d

Behind a Caddy reverse proxy

Expose the HTTP server on a hostname with automatic TLS. Add to your Caddyfile:

# Internal (self-signed) — homelab .example.invalid zone
jellyfin-mcp.example.invalid {
    tls internal
    reverse_proxy jellyfin-mcp-mcp:8000
}

# Public — automatic Let's Encrypt
jellyfin-mcp.example.com {
    reverse_proxy jellyfin-mcp-mcp:8000
}

Reload Caddy:

docker compose -f services/caddy/compose.yml exec caddy caddy reload --config /etc/caddy/Caddyfile

DNS with Technitium

Point the hostname at the host running Caddy. Via the Technitium API:

curl -s "http://technitium.example.invalid:5380/api/zones/records/add" \
  --data-urlencode "token=$TECHNITIUM_DNS_TOKEN" \
  --data-urlencode "domain=jellyfin-mcp.example.invalid" \
  --data-urlencode "zone=arpa" \
  --data-urlencode "type=A" \
  --data-urlencode "ipAddress=192.0.2.10" \
  --data-urlencode "ttl=3600"

…or add an A record jellyfin-mcp.example.invalid → <caddy-host-ip> in the Technitium web console (http://technitium.example.invalid:5380). The ecosystem technitium-dns-mcp automates this as a tool.

Register with an MCP client

Add to your client's mcp_config.json:

{
  "mcpServers": {
    "jellyfin-mcp": {
      "command": "uv",
      "args": ["run", "jellyfin-mcp"],
      "env": {
        "JELLYFIN_URL": "http://your-jellyfin:8096",
        "JELLYFIN_API_KEY": "your_api_key"
      }
    }
  }
}

For a remote HTTP server, point the client at http://jellyfin-mcp.example.invalid/mcp instead.