Modern social media systems are mostly designed toprioritise metrics based on engagement, which sometimes leads tounintentional promotion of social comparison, anxiety, loneliness,and emotional distress. While these systems usually matchusers based on surface-level factors such as similar interests,followers, or previous interactions, little attention is paid to thepsychological compatibility between them. This problem pointsout that current social systems still lack proper investigation andutilisation of emotionally relevant and supportive connections.This paper introduces MIND-CONNECT, a social compatibilityand emotional wellness system with personality modeling, intelli-gent recommendations, and LLM-driven conversation support.The system uses scientific research on human psychologicalcompatibility based on the Big Five (OCEAN) personalitymodel through a ten-question BFI-10 personality test providedduring user registration. Personality traits are mapped intoa five-dimensional vector space, and users’ compatibilities arecalculated using the Euclidean distance metric between theirpersonality vectors.In addition to compatibility-based social networking functional-ity, the system also implements an AI-based journaling and re-flective practice powered by the Groq API using the LLaMA 3.3-70B model. To enhance contextual relevance, the AI companioninjects information about the user from their own journal entries,allowing for emotionally aware interactions. The platform alsoprovides essential social networking services such as communities,posting and sharing content, sending direct messages, userdiscovery, and a role-based admin panel. The system is built asa web application consisting of React as its front-end, Express.jsback-end, and PostgreSQL database. Deployment takes place onVercel.The experimental testing shows the correct implementation ofpersonality scoring and compatibility calculation algorithms.Preliminary qualitative assessment suggests that the systemis capable of generating contextually coherent and empathicresponses from the AI companion. This proposal will contributeto creating future social media systems that focus more on thepsychological wellbeing and emotional connection between usersthan just increasing engagement.
goyal et al. (Sat,) studied this question.