In the framework of a collaborative research project that aims to understand, reproduce and analyze voice in healthy and pathological cases using virtual and robotic avatars, we present the latest advances on the development of a mechatronic testbed named “Pinocchio”. Reproducing in vitro the dynamic movement of speech articulators, such as jaw, tongue, velum and larynx, together with phonation, remains a challenge. Design efforts were focused on three aspects: how to breath, how to phonate, how to articulate. Air supply is controlled through a valve opening linearly related to airflow rate at the device inlet, a settling chamber stabilizing upstream jet, and a long tube leading from the chamber to the laryngeal model. Phonation is controlled by acting on a flexible laryngeal envelope and 1:1 replica of vocal-folds replica through lateral compression and longitudinal stretching. Depending on the choice of folds geometry and material tensile stiffness, self-oscillations generate a voice source covering a pitch range from less than 110 Hz to more than 220 Hz. Adduction and abduction articulatory movements are implemented within the larynx. Two primary articulatory movements are implemented within a replica of the vocal-tract : mandibular motion and shaping of the oral cavity with the movement of tongue body and apex. F1-F2 vocalic space produced by movements of jaw and tongue was measured. At this early stage of his life, Pinocchio is able to produce vowels in the surrounding of an [a], with F1 between 500 and 700Hz and F2 between 1000 and 2000Hz.