This contribution aims at (1) discussing the characteristics of collecting, filing and storing data to have a databank of oral interactions between university students whose main objective is the learning of a second language through teletandem; and (2) defining the steps for further collections and storage. Our data are Skype sessions of foreign language learners who interact via Voice Over Internet Protocol (VOIP) with a proficient partner in the language they are learning. Our databank aims at (1) giving value to teletandem as a situated learning context, (2) substantiating the research carried out in the field, and (3) offering other researchers the possibility to access data to confirm or refute published research. We first define a schema for interpreting teletandem sessions according to the Interaction Space (IS) Model as defined by Chanier and colleagues (2014). Subsequently, we discuss metadata concerning contexts (e.g. description of the university and of the language courses) and learning scenarios (e.g. objectives, materials). [For the complete volume, see ED571330.]