Apologies for the long video, but here’s a demo of our upcoming Publishers feature integrating with arbitrary lexicons. Integrating with ActivityPub is next, and mostly config. Since our protocol uses standard documents (usually webpages) as the canonical data source, this essentially lets you treat your web page as a protocol-agnostic data store. Our RDF substrate is much more rudimentary than yours, but the Harmonizer step allows you to define how you want to parse an existing document in an extensible manner, so theoretically you could use it to decompose a document into something as granular as you are looking for.