ABSTRACT

In recent years, university students have experienced a significant rise in mental health challenges, often feeling isolated, stressed, and overwhelmed. Traditional mental health services at universities are frequently limited by availability, long waiting times, and the stigma that discourages some students from seeking help. This increasing demand has underscored the urgent need for a scalable, accessible, and non-judgmental support system that offers timely guidance and mental health resources. To address these critical gaps, we developed an AI-powered chatbot specifically designed to support the mental health needs of students, offering them a supportive and confidential space to engage and seek assistance. Our solution was built using the MERN stack—a combination of MongoDB, Express.js, React, and Node.js—chosen for its robustness, scalability, and flexibility in handling both real-time interactions and data management. MongoDB, a NoSQL database, was selected for its ability to store and manage unstructured data, such as conversational logs and user entries, which are essential for enhancing the chatbot’s learning and responsiveness. Express.js and Node.js provide a powerful backend framework that ensures smooth communication between the server and client, enabling the chatbot to process requests quickly and maintain a seamless user experience. React was used for the front-end interface, creating a responsive, user-friendly platform that students can navigate easily, whether on mobile or desktop. Our project objectives included designing a reliable and interactive platform that provides a comprehensive range of mental health resources, including self-help guides, stress management strategies, and quick access to emergency contacts and counseling services. Additionally, we aimed to develop a system that could detect early signs of distress and escalate responses in crisis situations, facilitating timely referrals to professional support where necessary. We used an Agile methodology, iteratively enhancing the chatbot's capabilities based on continuous feedback from early testers and adapting to evolving mental health needs. This approach allowed us to refine the bot’s responses and add features that aligned more closely with students’ real-world challenges. At the core of the chatbot's functionality is advanced natural language processing (NLP), which enables it to engage in meaningful and empathetic conversations with users. Through the use of NLP models, the chatbot can understand context, respond appropriately, and detect sentiment, making interactions feel more supportive and tailored to individual needs. Ethical considerations, especially around privacy, confidentiality, and data security, were prioritized to ensure that students feel safe using the chatbot without concerns about data misuse. Results from initial testing indicate that the chatbot effectively meets many mental health support needs of students. Users reported that the platform was easy to use and appreciated its immediate availability, which helped reduce feelings of isolation and encouraged proactive engagement with mental health resources. By empowering students to manage stress, anxiety, and other mental health concerns independently, the chatbot shows promise in contributing to a healthier, more supportive campus environment. In sum, this chatbot provides an accessible, scalable solution that complements traditional mental health services. Its MERN stack foundation, combined with iterative Agile development and a strong focus on user privacy, makes it a reliable and adaptable resource for supporting the mental well-being of students within a higher education setting.