Some work will be needed for a complete mapping of the detailed information coming out of the API. Not only because the data is not always fully accurate (the API still thinks that Ian Davis and I are the same person) but also because the claims are document-oriented while most SG-related RDF vocabs are person-centric.
However, for any given URL somehow associated with a person, the API returns a set of identifiers that are very likely to lead to related data. So, for an RDF toolkit, these pointers are often already sufficient to send out its RDF extractors and enrich the local dataset. The SG API Parser that was now added to ARC (revision 2008-07-15) is still pretty basic, but it will generate rdfs:seeAlso triples for the canonical_mapping's value (as subject) and every mentioned HTTP identifier (as object).
I'm working on more low-level/direct RDF mappings for POSH formats such as XFN, those could simplify detailed triple extraction (w/o too much of the current person --homepage-> document indirection) from the API results.
Using the new parser in ARC is identical to working with any other syntax. The format detector will auto-include the necessary components. Just call