Between 2003-2005 I often used Swi-Prolog for semantic web experiments before I more or less settled on using Sesame (and occasionally AllegroGraph and Redland).
git clone git://eculture.cs.vu.nl/home/git/eculture/ClioPatria.git
Then, point your browser to http://localhost:3020/, load, for example, a OWL (XML format) data file, and try some queries. The default query language is SeRQL which I don’t use, so I set the query language to SPARQL and all seems to work fine.
One good thing about Swi-Prolog and the bundled SemWeb library is that loading RDF data and performing queries seems very quick compared to Sesame which is what I usually use. As a result, the ClioPatria web application is also very quick.