This article explores a number of possible corpus-based applications in translator training. The first involves the use of corpora created by translators (CCBT), a type of learner corpora that can be used to investigate difficulties encountered by trainee translators. The second focuses on the use of corpora created for translators (CCFT), monolingual target-language reference corpora than can be used as a resource for finding translation equivalents at a number of levels, including lexical, phraseological, syntactic and stylistic. Finally, comparable corpora (CC), which consist of translations into a given language alongside similar texts that have been originally written in that same language, are considered. These CC can be created by combining CCBT and CCFT, and they can be used as evaluation corpora to help trainers provide feedback on student translations, or as a body of data for investigating the nature of translated text as compared to original language text.