The Linked Data Fragments website provides a demo how some parts of your question can be solved. With Comunica you can do federated SPARQL queries from your browser using a variety of endpoints.
See : try the DBpedia,Harvard,VIAF example that was created do demonstrate how , for instance, libraries can use federated queries to match book authors against VIAF (a database of author names) and DBPedia (to provide more information authors).
We used this technique at Ghent University Library to enrich our catalog with global unique identifiers for book authors.
such as using the relationship between data across endpoints to query, revealing hidden relationships between endpoints, etc.
Comunica does have support for link traversal and there are some open challenges in SolidLab which look into being able to to use other endpoints (such as SPARQL endpoints and LDF) in the traversal process. These challenges are likely to be implemented as additional components for Comunica.
But it’s probably not what I’m looking for, it doesn’t seem to dig deeper into the semantics,
I’m not quite sure what you specifically mean by this; but if you are after a RDF reasoner then there are also Comunica components for that. We are hoping to improve the capabilities of that reasoner in the future using eye-js which has support for N3 Rules, RDF Surfaces and all of the other wonderful things that eye supports.