The past decade has seen a massive growth in online academic courses, most of which are offered in the English language. However, although more people speak English as their second language than as their first, online course providers do not offer language assistance. Providing online learners with language resources would allow them to both learn about a subject through a foreign language and learn the foreign language through the subject. This is referred to as “content-based language learning”. Supporting content-based language learning using online courses raises several challenges, three of which are addressed in this thesis. First, courses teach subjects in particular domains, but supporting domain-specific language requires knowledge of specialised vocabulary. This thesis develops an automated approach to generating domain-specific corpora and wordlists, extracting domain-specific vocabulary in a way that can be applied to any online course. Second, acquiring and measuring language come hand-in-hand. Tools that help learners acquire new language should also include methods for testing. This thesis takes an existing vocabulary test and automates it. This has two main advantages: it requires no assumed knowledge of the language, allowing automatic generation of vocabulary tests; and the tests reflect the wordlists used to create them, allowing them to be targeted toward a particular domain. Finally, for content-based language learning to be used successfully, the language components must be smoothly integrated into courses without disturbing the original content. Furthermore, vocabulary support should include multi-word lexical items as well as single words. The thesis describes a tool that enhances online course content, via a browser extension. It is completely automated, though would also lend itself to selective teacher intervention. It is illustrated here with reference to courses offered by the FutureLearn MOOC consortium.