This project is designed and implemented with the use of Computer Music Therapy (CMT) System, with multiple modalities to improve emotional well-being through adaptive music recommendations. The paper opens by surveying the existing literature about music recommendation and emotion-aware applications, pointing out that most of these platforms exploit users’ history and genre, but do not offer real-time emotional adaptation. Despite new emotion-based recommendations, those systems either focus on a single interaction modality or lack practical, user focused deployment suggesting integrated and accessible methods are in demand. This paper makes a proposal of a novel web-based music therapy system with the combination of three interactive modes such as chat-based text input, speech information input, and real-time facial emotion recognition to detect user’s emotions for providing appropriate music decision. This framework is designed to be modular, responsive and accessible and includes things like song language select menus, compatibility with web and mobile browsers.