This paper describes the design and construction of CHEU-lex, a parallel and comparable corpus of Swiss and European Union (EU) legislation. Data are available in the three languages of the Swiss Confederation (French, German and Italian) and include bilateral agreements between Switzerland and the EU and their reception in Swiss law. The corpus is a richly annotated multilingual resource and allows the analysis of legal language at several levels (macro-textual, lexical, morphosyntactic) and according to different perspectives (monolingual, cross-lingual, cross-textual, diachronic). The goal is to highlight key properties of CHEU-lex, discuss issues of legal corpus compilation and, finally, outline some applications for translation and legal linguistic research.