tdarr

Tdarr

Tdarr is a distributed media transcoding automation platform. It scans a media library, applies codec conversion rules (H.265, AV1, etc.), and manages a queue of transcode jobs across one or more worker nodes. The result is a smaller library with files in modern codecs — meaning less storage consumed and more Direct Play for Plex clients instead of on-the-fly transcoding.

Run it when:

• Your library has a lot of H.264 content that could shrink 40–60% by converting to H.265/HEVC
• You want Plex to Direct Play more often (fewer transcodes = less CPU/GPU load at watch time)
• You have idle GPU cycles that could be doing useful work overnight

Architecture

┌─────────────────────────────────────────────────┐
│ TrueNAS Docker Host                             │
│                                                  │
│  ┌──────────────┐    ┌──────────────────────┐   │
│  │ tdarr        │    │ tdarr-node           │   │
│  │ (server)     │◄───│ (TrueNAS GPU worker) │   │
│  │              │    │ network_mode:        │   │
│  │ :8265 webUI  │    │   service:tdarr      │   │
│  │ :8266 comms  │    │ NVIDIA GPU           │   │
│  └──────┬───────┘    └──────────────────────┘   │
│         │                                        │
│  /mnt/Data/Media (shared volume)                │
└─────────┼───────────────────────────────────────┘
          │ Traefik
          ▼
    tdarr.yourdomain.com (LAN only)

The server and node run in the same compose stack. The node uses network_mode: service:tdarr — it shares the server container's network namespace and communicates over loopback (serverIP=0.0.0.0), so no extra port exposure is needed for the node.

A second node on the gaming desktop connects to port 8266 over the LAN — see ../tdarr-desktop-node/.

GPU contention note: The TrueNAS GPU is also used by Plex. transcodegpuWorkers=1 caps Tdarr to one GPU encode job at a time, leaving headroom for Plex streams. If you notice Plex transcoding degradation during heavy Tdarr runs, drop this to 0 during peak hours or use Tdarr's built-in scheduler.

Prerequisites

TrueNAS host:

• NVIDIA Container Toolkit installed (nvidia-ctk available)
• Docker daemon configured with the NVIDIA runtime
• Host directories created before first run:

  mkdir -p /mnt/SSD/Containers/tdarr/{server,configs,logs,transcode-cache}
  

Quick Start

cp example.env .env
# Edit .env — set TDARR_DOMAIN
docker compose up -d

Open https://${TDARR_DOMAIN} → Libraries → add a library pointing to /data/media.

Environment Variables

Variable	Default	Description
TZ	America/New_York	Container timezone
PUID	3001	User ID for file ownership
PGID	3001	Group ID for file ownership
TDARR_DOMAIN	—	Traefik domain for the web UI
TDARR_PORT	8265	Host port override (optional)

Ports

Port	Purpose
8265	Web UI (exposed via Traefik)
8266	Node communication (direct, LAN only)

Volumes

Host Path	Container Path	Purpose
/mnt/SSD/Containers/tdarr/server	/app/server	Server queue database
/mnt/SSD/Containers/tdarr/configs	/app/configs	Shared node configs
/mnt/SSD/Containers/tdarr/logs	/app/logs	Shared logs
/mnt/Data/Media	/data/media	Media library (read/write)
/mnt/SSD/Containers/tdarr/transcode-cache	/temp	Transcode scratch space