An Internet-based media service platform can control recording processes and manage video and audio data. Furthermore, the design and implementation of an object-based system for recording enable the flexible playback of the viewing contents. Augmented Reality (AR) is a three-dimensional video projection technology. However, there are few examples of its use as a method for audiovisual media platforms. In this study, we propose Co-Sound, which is designed as a multimodal interface that renders object-based AR dynamically in response to various actions from viewers on a web browser by sharing AR objects among multiple devices in real time. We confirmed that the system was developed as an object-based interactive medium with AR, achieved the general acceptance of the system was very high through a questionnaire survey, and low-latency synchronization to accept operations from multiple users in real time.