<p>Metaverse is a digitally created universe, where users are able to interact in real time with other user in a computer-generated environment. In metaverse applications users are represented through a digital identity known as avatar. Through these avatars, communications are established and application’s objectives are achieved. At present, there are many social-media applications that makes use of these avatars as the digital identities. These avatars are static in nature and are predefined. However, there is gap in the state-of-the-art avatar generators, where the present emotional state of the user is not considered for avatar generation. In this regards, to address this challenge, we propose a generative artificial intelligence framework named “DynamoAvatar” for avatar generation. It utilizes both static information and dynamic behavior of the user to generate real-time dynamic avatars of the users in metaverse. Our experimental results demonstrate an accuracy of 92% for dynamic avatar generation with improved user acceptance rate. These results are promising for improving the adoption rate of metaverse applications by the user community.</p>