In mental disorders, the diagnosis of depression has become a big challenge for psychiatrists as well as for professionals. Clinical treatment for depression is a heavy task, therefore, there has been renewed interest in using artificial intelligence approaches for recognizing depression clues based on machine learning algorithms and natural language processing techniques. This paper introduces a machine learning-based chatbot application for achieving two objectives with depression as a mental disorder type, firstly, achieving awareness of various kinds of mental disorders based on automatic answers to multiple questions about mental disorders provided by users to the proposed chatbot. Secondly, recognizing depression symptoms, which are useful to diagnose depression cases. The performance of the proposed chatbot app has been tested and evaluated using three machine learning algorithms, logistic regression algorithm, linear SVM, and neural networks to diagnose depression cases. The performance of three algorithms has been tested and compared on a benchmark dataset. The testing accuracy results clarified the superiority of the logistic regression algorithm and neural networks over linear SVM in diagnosing depression cases with a testing accuracy of 91%.