Great post. The gap I see is a mechanism like a dropbox where you can select or drag audio files to for ingestion into a pipeline for identification using beets (https://beets.io/) or MusicBrainz Picard, where they’re autoclassified, and then uploaded into the storage system that underlies the media server.

thanks! I have samba running on my dietpi. When picard saves -> staging folder -> dietpi. done. EDIT: not fully automated yet, because picard needs manual intervention at times (classification mix-ups), but pretty pain-free overall.