The document describes a proposed system for building a chatbot that takes gesture-based user input. The system would use either a deep learning model or convexity defect algorithm to recognize gestures from video input. Recognized gestures would be mapped to text commands and fed into a keyword-based chatbot. The chatbot would execute commands or responses based on the gesture input. The proposed system aims to provide a natural interface for applications helping deaf/mute users or in places like museums. It reviews related work on gesture recognition and discusses the technical components and workflow of the envisioned chatbot system.