Recommender systems have become a fundamental tool for enhancing user experiences in various contexts, such as social networks, e-Commerce, streaming platforms, etc. No one doubts the capabilities of these types of tools, but the personalized demands of users are growing significantly, generating an immediate effect on the development of more complex systems. From this perspective, we have developed an application called Swappito ( www.swappito.com ) contextualized in the field of exchanges. This application aims to create a sustainable ecosystem for the exchange of second-hand products and skills among users, where the main concept introduced is bartering. It is in this context wherein we focus on the work of this proposal; i.e. to develop a recommendation system for this application, which has the peculiarity that the relationship must be established between a pair of user and object with another pair of user and object, rather than the more traditional direct recommendation. Recent experiments carried out on content-based, collaborative filtering and hybrid algorithms have yielded promising results, providing information on the potential methodologies for such systems.