The rigorous education curriculum in Singapore and the local “kiasu” culture have many impacts on youths (aged 13 to 18). One of them is the increasing levels of stress that youths face, which may lead them to suffer from mental illnesses. The rising rates of mental illnesses such as depression in youths in Singapore have also led to a higher suicide rate in youths. Hence, it is important to have an accessible platform for youths to express their feelings as a form of therapy. By writing Python code which encompasses deep learning neural networks, a chatbot will be created on an online chatting platform, Telegram. It uses deep learning and semantic analysis to carry out a conversation with the user and detect his/her levels of negative and positive emotions. The code makes use of API (Application Program Interface) TextBlob to measure the levels of stress that the user is facing through an engaging conversation via semantic analysis. This is done by calculating the frequency of positive or negative words used in the user’s text messages. Besides being a detector of levels of stress in the user, our chatbot also has the potential to aid in online supportive therapy.