An eye-tracking study compared the effects of actions (depicted as tools between on-screen characters) with those of a speaker’s gaze and head shift between the same two characters. In previous research, each of these cues has rapidly influenced language comprehension on its own, but few studies have directly compared these two cues or, more generally, distinct non-linguistic cues in their effects on realtime sentence comprehension. We investigated how participants used action tools and speaker gaze separately and in combination for visually anticipating the upcoming mention of a sentence referent. We discuss implications for accounts of visually situated language comprehension.