This study investigates how learners use a freely available web-based multimodal language learning environment based on videos and newspaper articles with corresponding exercises and learning support in the form of help options, "NedBox," for learning L2 Dutch. In order to measure learners' use, a combination of eye movement and screen recording data were analysed. The study also examines how learners' eye movements and use of the environment are related to their proficiency and performance in the environment. Twenty-one learners of Dutch used NedBox for 40 minutes, while their eye movements were tracked and their screen was recorded. Results demonstrated that learners used the environment in different ways in order to deal with the simultaneous presentation of input and exercises, and that this was dependent on the type of input. Eye movements were, to some extent, related to proficiency and performance in the environment. Finally, learners' use of the help options can be defined as idiosyncratic. In sum, the findings of this study suggest that the simultaneous presentation of input, be it video or text, does not overwhelm learners and that learners do not always seem to use help options effectively.