This paper presents the design and implementation of a novel social media application where internal data is structured as a hypergraph. Users are represented as hypernodes, and topics are modeled as hyperedges that encircle the hypernodes. Hypergraphs allow for modeling complex relationships, making them an alternative to traditional graph-based approaches. The application leverages modern web development technologies to provide a seamless user experience. A Python backend integrates machine learning models to extract topics (mentioned as hyperedges here) from posts and update the hypergraph database dynamically. The backend was tested with various models, and the combination of TF-IDF and Naive Bayes was ultimately chosen for its superior accuracy. This study highlights a hypergraph-based representation of a social network and discusses the practical challenges encountered during implementation.