In this paper, we apply Semantic Web technologies to the creation of an improved search engine over legal and public administration documents. Conventional search strategies based on syntactic matching of tokens offer little help when the users' vocabulary and the documents' vocabulary differ. This is often the case in public administration documents. We present a semantic search tool that fills this gap using Semantic Web technologies, in particular, ontologies and controlled vocabularies, and a hybrid search approach, avoiding the expensive tagging of documents.