Tracheal intubation is considered the gold standard to secure the airway of patients in need of respiratory assistance, yet this procedure relies on the dexterity and experience of the physician to correctly place a tracheal tube into the patient's trachea. Such a complex procedure may greatly benefit from robotic assistance in order to make the intubation safer and more efficient. We developed the first device to provide such assistance, the REALITI, which stands for Robotic Endoscope Automated via Laryngeal Imaging for Tracheal Intubation. This device allies the automated detection of key anatomical features in an endoscopic image to the robotic steering toward the recognized features in the task of guiding the tracheal tube into its correct position. The pre-clinical prototype presented in this paper has been developed to perform in vitro tracheal inbutation on a standard airway management training manikin. We performed a robust detection of anatomical features to steer the endoscope in a visual servoing fashion. Our prototype has been successfully used to perform automated and manual insertions into the trachea of an airway manikin.