Intracranial human recordings are a valuable and rare resource that the whole neuroscience community can benefit from. Making such data available to the neuroscience community not only helps tackle the reproducibility issues in science, it also helps make more use of this valuable data. The latter is especially true for data collected using naturalistic tasks. Here, we describe a dataset collected from a large group of human subjects while they watched a short audiovisual film. The dataset is characterized by several unique features. First, it combines a large amount of intracranial data from 51 intracranial electroencephalography (iEEG) participants, who all did the same task. Second, the intracranial data are accompanied by fMRI recordings acquired for the same task in 30 functional magnetic resonance imaging (fMRI) participants. Third, the data were acquired using a rich audiovisual stimulus, for which we provide detailed speech and video annotations. This multimodal dataset can be used to address questions about neural mechanisms of multimodal perception and language comprehension as well as the nature of the neural signal acquired during the same task across brain recording modalities.