Although online tutorials are becoming commonplace for language teaching, very few studies to date have provided insights into learners' behaviours in synchronous online interactions from their own perspective. This study employs eyetracking technology to investigate ten learners' attention during synchronous online language learning in a multimodal environment. The participants were learners of Chinese as a Foreign Language at beginner's or lower-intermediate level. While learners took part in two different online activities, one focusing on reading, the other on interaction with others, their gaze focus was tracked, and in subsequent stimulated recall interviews the learners reflected on their engagement with the screen and their intentions while reading or speaking online. Our findings show that during reading tasks, when Pinyin transcriptions as well as Chinese characters were presented, all beginner and lower intermediate participants focused to some degree on the Pinyin. In the interactive task learners' gaze was drawn to elements of the screen that were not immediately necessary for technical or linguistic reasons but that could be interpreted as containing social presence information, e.g. names listed and emoticons employed by other users.