The paper describes the principle of organizing the multimodal component of the Corpus
“Konduit” — the corpus of child oral unprepared elicited narratives. The corpus consists of 213 oral
texts produced by Russian-speaking children aged 2, 7–7, 6 years old and presented in the form of
annotated texts, as well as annotated and synchronized audio and video files. The “Konduit” Corpus
is currently the only Russian-language corpus of oral children's texts with multimodal data in the
public domain.