To explore the possibility of conducting online experiments on audiovisual crossmodal interaction, we investigated the temporal accuracy and precision of stimulus presentation in Lab.js. We measured the presentation time as well as the onset and offset timings of audiovisual stimuli in two experiments; we used three operating systems (Windows10, macOS, and iOS), three browsers (Chrome, Firefox, and OS native), and three network environments (external network, internal network, and local host). Results showed that most software environments—with the exception of iOS—provide reasonable precision, irrespective of the network environment. In contrast, the accuracy almost systematically varied depending on the combination of the operating system and the browser. These results suggest that Lab.js is suitable for audiovisual experiments to some extent if audiovisual synchronization lag is corrected based on the information provided by the Lab.js default plug-in, which can detect the user’s OS and browser. However, considering that typical user hardware/software environments are far more complex and diverse than that of this study, effective ways to control users’ experimental environments should be investigated in future studies to obtain reliable data in online experiments.