The hardware I’ve set up for Emby isn’t particularly remarkable. It’s my old daily driver: an Intel i7-7700 paired with 48GB of RAM and an ageing GTX 1660 GPU handling the transcoding duties.
For the operating system, I opted for Windows Server 2025, as I happened to have an unused licence key available. It’s installed on its own dedicated 64GB NVMe drive with the paging file disabled.
Emby itself is installed on the system drive, while the cache and image storage are handled by a second NVMe drive (500GB), installed using a PCIe x16 to NVMe adapter.
My media library is quite extensive. It includes music, audiobooks, ebooks, films, home videos, television series, photo uploads and YouTube downloads. This main library, not including the comic book collection (which spans 50TB and resides on a separate instance), holds roughly 3.3 million entries. These are spread across 50TB of local RAID volumes for frequently accessed content, and four Synology NAS systems for archived material, totalling approximately 400TB of raw capacity.
Films are organised by decade, distributed across multiple volumes and disk groups. TV series are divided into two separate libraries: one for Running series and another for Archived or no longer running stuff. Music, ebooks and audiobooks are housed on a separate local volume.
For conversion and transcoding tasks, I use older SATA SSDs configured in RAID 0. As this is a non-critical function, I’m not concerned about potential drive failures in this setup.
As clients I have a Nvidia Shield TV 2019, a Nokia Box 8000, an Apple TV 4, 2 x 1080p Android beamers, 3 Smart TVs (LG and Samsung), various iPads, iPhones and Android phones/tablets on various OS versions that connect regularly from various counties mainly within Europe.