We have developed a False-Belief (FB) understanding task for use on a computer tablet, trying to assess FB understanding in a less social way. It is based on classical FB protocols, and additionally includes a manipulation of language in an attempt to explore the facilitating effect of linguistic support during FB processing. Specifically, the FB task was presented in three auditory conditions: narrative, silent, and interference. The task was assumed to shed new light on the FB difficulties often observed in Autism Spectrum Disorder (ASD). Sixty-eight children with ASD (M = 7.5 years) and an age matched comparison group with 98 typically developing (TD) children were assessed with the FB task. The children with ASD did not perform above chance level in any condition, and significant differences in success rates were found between the groups in two conditions (silent and narrative), with TD children performing better. We discuss implications, limitations, and further developments.