Deployment¶

Deployment Options¶

arr-mcp exposes its MCP server (console script arr-mcp) four ways. Pick the row that matches where the server runs relative to your MCP client, then copy the matching mcp_config.json below. Replace the <your-…> placeholders with the values from the Configuration / Environment Variables section.

#	Option	Transport	Where it runs	`mcp_config.json` key
1	stdio	`stdio`	client launches a subprocess	`command`
2	Streamable-HTTP (local)	`streamable-http`	a local network port	`command` or `url`
3	Local container / uv	`stdio` or `streamable-http`	Docker / Podman / uv on this host	`command` or `url`
4	Remote URL	`streamable-http`	a remote host behind Caddy	`url`
5	Open WebUI	`streamable-http`	any reachable HTTP endpoint	`url`

1. stdio (local subprocess)¶

The client launches the server over stdio via uvx — best for local IDEs (Cursor, Claude Desktop, VS Code):

{
  "mcpServers": {
    "arr-mcp": {
      "command": "uvx",
      "args": ["--from", "arr-mcp", "arr-mcp"],
      "env": {
        "ARR_HOST": "<your-arr_host>"
      }
    }
  }
}

2. Streamable-HTTP (local process)¶

Run the server as a long-lived HTTP process:

uvx --from arr-mcp arr-mcp --transport streamable-http --host 0.0.0.0 --port 8000
curl -s http://localhost:8000/health        # {"status":"OK"}

Then either let the client launch it:

{
  "mcpServers": {
    "arr-mcp": {
      "command": "uvx",
      "args": ["--from", "arr-mcp", "arr-mcp", "--transport", "streamable-http", "--port", "8000"],
      "env": {
        "TRANSPORT": "streamable-http",
        "HOST": "0.0.0.0",
        "PORT": "8000",
        "ARR_HOST": "<your-arr_host>"
      }
    }
  }
}

…or connect to the already-running process by URL:

{
  "mcpServers": {
    "arr-mcp": { "url": "http://localhost:8000/mcp" }
  }
}

3. Local container / uv¶

(a) Launch a container directly from mcp_config.json (stdio over the container — no ports to manage). Swap docker for podman for a daemonless runtime:

{
  "mcpServers": {
    "arr-mcp": {
      "command": "docker",
      "args": [
        "run", "-i", "--rm",
        "-e", "TRANSPORT=stdio",
        "-e", "ARR_HOST=<your-arr_host>",
        "knucklessg1/arr-mcp:latest"
      ]
    }
  }
}

(b) Run a local streamable-http container, then connect by URL:

docker run -d --name arr-mcp -p 8000:8000 \
  -e TRANSPORT=streamable-http \
  -e PORT=8000 \
  -e ARR_HOST="<your-arr_host>" \
  knucklessg1/arr-mcp:latest
# or, from a clone of this repo:
docker compose -f docker/mcp.compose.yml up -d

{
  "mcpServers": {
    "arr-mcp": { "url": "http://localhost:8000/mcp" }
  }
}

(c) From a local checkout with uv:

uv run arr-mcp --transport streamable-http --port 8000

4. Remote URL (deployed behind Caddy)¶

When the server is deployed remotely (e.g. as a Docker service) and published through Caddy on the internal *.arpa zone, connect with the "url" key — no local process or image required:

{
  "mcpServers": {
    "arr-mcp": { "url": "http://arr-mcp.arpa/mcp" }
  }
}

Caddy reverse-proxies http://arr-mcp.arpa to the container's :8000 streamable-http listener; http://arr-mcp.arpa/health returns {"status":"OK"} when the service is live.

5. Open WebUI¶

Open WebUI connects to MCP servers via their Streamable-HTTP endpoint. Run arr-mcp with the streamable-http transport so Open WebUI can reach it by URL:

arr-mcp --transport streamable-http --host 0.0.0.0 --port 8000

Or with Docker, ensuring the port is published:

docker run -d --name arr-mcp -p 8000:8000 \
  -e TRANSPORT=streamable-http \
  -e PORT=8000 \
  -e ARR_HOST="<your-arr_host>" \
  knucklessg1/arr-mcp:latest

In Open WebUI, navigate to Workspace → MCP Servers → Add Server and enter:

Field	Value
Name	`arr-mcp`
Server URL	`http://<host>:8000/mcp`

Replace <host> with the address where arr-mcp is reachable from Open WebUI. When running on the same Docker network or host, http://localhost:8000/mcp or http://arr-mcp:8000/mcp (Docker service name) work as the URL.

No additional headers are required. Open WebUI discovers the full tool set (radarr_action, sonarr_action, lidarr_action, etc.) automatically via the MCP protocol's tools/list handshake.

To test the connection, use Open WebUI's built-in tool call editor or send a simple query like "Show me movies in Radarr" — the LLM will call radarr_action with action get_movie to list the catalog.

Action naming: All actions follow the API client method names. Use action='list_actions' to discover every valid action for a service. Common plurals like get_movies are aliased automatically to get_movie by the dispatch resolver. A did-you-mean suggestion is returned for any action name that can't be resolved.

This page covers running arr-mcp as long-lived servers: the MCP transports, the companion A2A agent, a Docker Compose stack, putting it behind a Caddy reverse proxy, and giving it a DNS name with Technitium. To provision the Arr Suite services it connects to, see Backing Platform.

arr-mcp ships two console scripts: an MCP server (arr-mcp) and an A2A agent server (arr-agent). The MCP server is a typed, deterministic tool surface; the agent server is a Pydantic-AI agent that calls those tools over the Agent Control Protocol.

Run the MCP server¶

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

stdio (default)streamable-httpsse

arr-mcp

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

arr-mcp --transport streamable-http --host 0.0.0.0 --port 8000

A network server with a /health endpoint and /mcp route.

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

Health check (HTTP transports):

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

Configuration (environment)¶

arr-mcp is configured entirely from the environment. The server-level settings:

Var	Default	Meaning
`HOST`	`0.0.0.0`	Bind address (HTTP transports)
`PORT`	`8000`	Listen port (HTTP transports)
`TRANSPORT`	`stdio`	`stdio`, `streamable-http`, or `sse`
`ENABLE_OTEL`	`True`	OpenTelemetry / Langfuse export
`EUNOMIA_TYPE`	`none`	Authorization mode: `none`, `embedded`, `remote`

Each Arr service is connected with its own block; a connector remains inactive when its credentials are absent:

Var	Example	Meaning
`SONARR_BASE_URL`	`http://localhost:8989`	Sonarr base URL
`SONARR_TOKEN`	`your_sonarr_api_key`	Sonarr API key
`RADARR_BASE_URL`	`http://localhost:7878`	Radarr base URL
`RADARR_TOKEN`	`your_radarr_api_key`	Radarr API key
`LIDARR_BASE_URL`	`http://localhost:8686`	Lidarr base URL
`LIDARR_TOKEN`	`your_lidarr_api_key`	Lidarr API key
`PROWLARR_BASE_URL`	`http://localhost:9696`	Prowlarr base URL
`PROWLARR_TOKEN`	`your_prowlarr_api_key`	Prowlarr API key
`BAZARR_BASE_URL`	`http://localhost:6767`	Bazarr base URL
`BAZARR_API_KEY`	`your_bazarr_api_key`	Bazarr API key
`SEERR_BASE_URL`	`http://localhost:5055`	Seerr base URL
`SEERR_API_KEY`	`your_seerr_api_key`	Seerr API key
`CHAPTARR_BASE_URL`	`http://localhost:8006`	Chaptarr base URL
`CHAPTARR_TOKEN`	`your_chaptarr_api_key`	Chaptarr API key

Each service also accepts a *_SSL_VERIFY flag (default False for self-signed homelab certificates). The full set, grouped by service, is documented in .env.example. Copy it to .env and populate only the services you use.

Docker Compose¶

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

services:
  arr-mcp-mcp:
    image: knucklessg1/arr-mcp:latest
    container_name: arr-mcp-mcp
    hostname: arr-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 the SONARR_* / RADARR_* values
docker compose -f docker/mcp.compose.yml up -d
docker compose -f docker/mcp.compose.yml logs -f

Run the A2A agent server¶

arr-mcp ships a second console script, arr-agent, a Pydantic-AI agent that calls the MCP tools over the Agent Control Protocol and exposes an optional web interface. It connects to the MCP server via MCP_URL and listens on :9099 by default. The repo ships docker/agent.compose.yml, which runs both servers together:

services:
  arr-mcp-mcp:
    image: knucklessg1/arr-mcp:latest
    container_name: arr-mcp-mcp
    hostname: arr-mcp-mcp
    restart: always
    env_file:
      - ../.env
    environment:
      - PYTHONUNBUFFERED=1
      - HOST=0.0.0.0
      - PORT=8000
      - TRANSPORT=streamable-http
    ports:
      - "8000:8000"

  arr-mcp-agent:
    image: knucklessg1/arr-mcp:latest
    container_name: arr-mcp-agent
    hostname: arr-mcp-agent
    restart: always
    depends_on:
      - arr-mcp-mcp
    env_file:
      - ../.env
    command: ["arr-agent"]
    environment:
      - PYTHONUNBUFFERED=1
      - HOST=0.0.0.0
      - PORT=9099
      - MCP_URL=http://arr-mcp-mcp:8000/mcp
      - PROVIDER=${PROVIDER:-openai}
      - MODEL_ID=${MODEL_ID:-gpt-4o}
      - ENABLE_WEB_UI=True
    ports:
      - "9099:9099"

docker compose -f docker/agent.compose.yml up -d
curl -s http://localhost:9099/health         # agent health

Var	Default	Meaning
`MCP_URL`	`http://arr-mcp-mcp:8000/mcp`	MCP server the agent calls
`PROVIDER`	`openai`	LLM provider
`MODEL_ID`	`gpt-4o`	Model identifier
`ENABLE_WEB_UI`	`True`	Serve the AG-UI web interface

Behind a Caddy reverse proxy¶

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

# Internal (self-signed) — homelab .arpa zone
arr-mcp.arpa {
    tls internal
    reverse_proxy arr-mcp-mcp:8000
}

# Public — automatic Let's Encrypt
arr-mcp.example.com {
    reverse_proxy arr-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.arpa:5380/api/zones/records/add" \
  --data-urlencode "token=$TECHNITIUM_DNS_TOKEN" \
  --data-urlencode "domain=arr-mcp.arpa" \
  --data-urlencode "zone=arpa" \
  --data-urlencode "type=A" \
  --data-urlencode "ipAddress=10.0.0.10" \
  --data-urlencode "ttl=3600"

…or add an A record arr-mcp.arpa → <caddy-host-ip> in the Technitium web console (http://technitium.arpa: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": {
    "arr-mcp": {
      "command": "uv",
      "args": ["run", "arr-mcp"],
      "env": {
        "SONARR_BASE_URL": "http://your-sonarr:8989",
        "SONARR_TOKEN": "your_sonarr_api_key",
        "RADARR_BASE_URL": "http://your-radarr:7878",
        "RADARR_TOKEN": "your_radarr_api_key"
      }
    }
  }
}

For a remote HTTP server, point the client at http://arr-mcp.arpa/mcp instead.