In RAM, we create hundreds of database indexes for the various types of metadata we hold, and cross references as well. One example of what this enables is that we can highlight artists in a bio/review that only exist in your own library.
For the content itself, it’s spread out and quite a bit larger than you’d expect. The typical compilation album has thousands of credits associated with it.
If you run on an SSD, you get a 10x-100x performance for jumping around the databases on disk (since we cant hold it all in RAM) – if you run on a spinning disk, it will work, but you will pay the penalty.
With a small number of albums, this will work fine… but as you grow your library by adding files or starring stuff in TIDAL, you will feel the pain.
Oh how I wish this was true 
Credits kill everything.
good! Glad you are happy with the experience. I have 12k albums and I was running in this exact configuration for a few weeks to debug some issues with performance… Everything was painfully slow.
I wouldn’t hold my breath for it, but its the most important thing on our roadmap, and probably a big part of our proper mobile solution.