What is this?
When you write an RDF-Connect processor definition you describe its configurable properties using SHACL shapes — because that's what the framework uses at runtime. You don't think about OWL semantics: property domains, ranges, or class declarations feel like unnecessary boilerplate there.
But LSP servers (like the Semantic Web Language Server) do need those OWL triples to give you good autocompletion and type-checking in your editor. This server bridges the gap: POST your definition and get back the OWL statements that can be inferred from it — without touching the original file.
Use it at build time
Pipe your definition through /reason and append the result to enrich the file:
What gets inferred?
sh:property with a sh:path becomes an owl:ObjectProperty or owl:DatatypeProperty with proper rdfs:domain and rdfs:range.sh:targetClass is declared as an owl:Class so editors recognise it as a type.sh:name are forwarded as rdfs:label on the property.