Abstract Program which is fueled by artificial intelligence to play out specific undertakings without the assistance of people is known as a bot. Chatbots are the most successive models of bots and these can make discussions with clients and they are flexible simultaneously. A Medical chatbot empower the movement of a clinical administrations provider and improves their presentation by partner with customers in human-like way. It is fundamental to observe that despite the way that chatbots can offer significant real factors and results, they can't give an official examination. In this paper, we created a chatbot which make conversation so that we can get the more information from the user and current state of mind. Our model can detect seven types of emotions from users’ inputs. For feeling distinguishing proof, we passed on two significant deep learning classifiers, Recurrent Neural Network (RNN), and Long-Short-Term-Memory (LSTM) and pre-trained weighted word index known as glove2. For better model getting ready and to avoid the overfitting of the model we applied hyperparameter tuning. The results were observable where the training accuracy of 88% and testing accuracy of 84%.Specifically, the proposed strategy of the chatbot is space explicit where through the clients’ connection, the chatbot will attempt to forestall the skeptical activities and revamp more useful considerations.