Purpose: Unmanic is the content transcoding and media optimization engine in your automation ecosystem. Its primary goal is to convert incoming media files into formats that balance efficient resource utilization and high-quality playback across devices, while integrating seamlessly with your acquisition pipeline (Prowlarr, Jackett, Sonarr, Radarr, qbittorrent).

Installed in a Python virtual environment on your bare-metal Plex Media Server, Unmanic handles CPU/GPU-aware video and audio transcoding, subtitle extraction, and file management, ensuring that your library stays consistent and optimized with minimal manual intervention.

—

• Plugins Stack – Unmanic operates as a sequence of plugins applied in processing order. Plugins can manage file acceptance, transcoding, subtitle handling, and post-processing tasks.
• Hardware Utilization – Unmanic can leverage NVIDIA GPUs (`nvenc`), Intel QuickSync (`qsv`), or CPU-based transcoding. Optimal configuration balances CPU/GPU usage with quality and throughput.
• Library Safety – Built-in filters prevent files from being unnecessarily re-encoded, reduce duplication, and reject output larger than the original.
• Subtitle Handling – All ASS or embedded subtitles are converted to SRT, and unwanted image-based subtitle streams are removed to maximize compatibility.

—

These plugins determine which files are processed and how:

• Ignore Recently Modified – Skip files modified in the last hour to avoid partial or actively written files.
• Ignore Under Size – Minimum file size threshold of 300 MB prevents transcoding very small clips.
• Ignore Video File Under Resolution – Default target 720p, ignoring 480p/SDTV content unless explicitly required.
• Reject File if Larger than Original – Prevents unnecessary bloating.
• Subtitle Extraction & Conversion – All ASS subtitles converted to SRT; image subtitles stripped or removed.
• Stream Stripping – Remove extraneous image streams and non-essential audio/subtitle languages to optimize file size.

Processing plugins handle actual transcoding and audio/video handling:

• Video Transcoding – Default codec: `hevc_nvenc` with CRF-style quality targeting. GPU `nvenc` device 0 used, preset `p6`, tune `hq`, main profile `main`. Smart filters applied, black bar autocropping enabled. Target resolution `source` (keeps original unless otherwise specified).
• Audio Transcoding – Main codec `aac` via standard VBR or AC3 encoder as fallback. Average bitrate 192 kbps; stereo fallback audio added to ensure device compatibility.
• Stream Cleanup – Remove unneeded image streams, unnecessary subtitle languages, and extra audio tracks.
• Container & Output – Output container: MKV; original container kept only if no transcoding applied.

Final checks and library integration:

• Reject File if Larger than Original – Last safeguard.
• Notifications – Radarr, Sonarr, and Plex are notified on task success or failure.
• Subtitle & Transcode Completion – Ensure all SRT subtitles are properly extracted; verify audio/video transcodes completed.

—

• Video Transcoder (`video_transcoder`)
  1. `mode`: standard
  2. `video_encoder`: `hevc_nvenc` (GPU 0)
  3. `preset`: 4 (`p6` for NVENC)
  4. `tune`: `hq`
  5. `CRF` based quality: 23
  6. Smart filters enabled
  7. Autocrop black bars enabled
  8. Target resolution: source
  9. Strip extra streams and attachments

• Audio Transcoder (`audio_transcoder`)
  1. Codec: AAC (VBR, 192 kbps)
  2. AC3 fallback for compatibility
  3. Extra stereo track added where applicable

• Subtitle Handling
  1. Extract all SRTs, convert any ASS to SRT
  2. Remove image-based subtitles

• Filters
  1. Ignore files under 300 MB
  2. Ignore files modified <1h ago
  3. Ignore video under 720p by default
  4. Remove unwanted subtitle languages (all non-English / non-primary languages)

—

• CPU – Multithreaded CPU usage is supported; efficient for smaller transcodes.
• GPU – NVIDIA NVENC preferred for HEVC hardware acceleration; ensures high throughput without overloading CPU.
• RAM – Recommended minimum 16 GB for a moderate library, more for concurrent transcodes.
• Disk I/O – SSD/NVMe recommended for active working folders; NFS mounts acceptable but slower.
• Python Virtual Environment – Ensures clean dependency management without affecting system Python.

—

• Start with a small test folder to ensure your hardware, GPU, and filters behave as expected.
• Always test Fallback AAC/AC3 audio to ensure device compatibility before bulk processing.
• Monitor logs for skipped or rejected files; this often indicates file size, resolution, or filter mismatches.
• Update Unmanic regularly; plugin improvements and bug fixes are frequent.

—

This guide is designed to get a new user running quickly with our optimized configuration, balancing high quality and efficient resource use while minimizing library disruption. Detailed plugin tuning or per-file troubleshooting should be referenced in the official plugin documentation as needed.