pal9000/shrinkerr repository overview

⁠Shrinkerr

Self-hosted media library transcoder with hardware encoding, VMAF quality assurance, and a safety net for your originals.

Re-encode H.264 / MPEG-2 / VC-1 / VP9 sources to HEVC, saving 40–65% disk space on a typical library. NVENC (NVIDIA) and libx265 (CPU) encoders; NVDEC / QSV / VAAPI hardware decode. VMAF auto-rejects bad encodes. Plex / Jellyfin / Emby aware (pauses on stream, refreshes libraries on completion). Sonarr / Radarr / NZBGet / SABnzbd integration with a rules engine. Distributed encoding across multiple workers.

⁠Image variants

All variants share the same DB schema and settings format — switch between them by changing one image: line and pulling.

⁠Quick start

# docker-compose.yml
services:
  shrinkerr:
    image: pal9000/shrinkerr:latest
    container_name: shrinkerr
    ports:
      - "6680:6680"
    volumes:
      - ./data:/app/data
      - /path/to/your/media:/media
    restart: unless-stopped

docker compose up -d

Then open http://localhost:6680⁠, add your media directories in Settings, and run a scan.

For NVENC, follow the GPU setup guide⁠ — needs the NVIDIA Container Toolkit on the host.

⁠Links

• Source + full documentation — github.com/I-IAL9000/shrinkerr⁠
• Installation guides (Portainer, reverse proxy, Windows + WSL2) — docs/installation.md⁠
• Encoder tuning, VMAF, hardware decode — docs/encoding-guide.md⁠
• Issues / feature requests — github.com/I-IAL9000/shrinkerr/issues⁠
• Changelog — CHANGELOG.md⁠

⁠Mirrored on GHCR

Also available without Docker Hub's anonymous-pull rate limits (100 / 6h / IP):

docker pull ghcr.io/i-ial9000/shrinkerr:latest

Same images, both registries published simultaneously from CI.