The purpose of our daydreaming model is to represent or "tell" computer generated daydreams employing computational animations. We employ the engagement part of the engagement-reflection model of writing to generate sequences of actions which are then animated employing computers. An important characteristic of our system is that animated characters are represented by non-verbal, non-anthropomorphic and non-zoomorphic visual objects. We refer to this program as the Visual-Daydreamer (V-Daydreamer). This paper reports on our first prototype. It describes some initial experiments that attempt to analyse if the outputs generated by the system are considered as daydreaming narratives by a group of human referees. The results suggest that V-Daydreamer's output corresponds with what most subjects identify as a daydream.