ByteDance has launched SeedRealtime, a native audio-video full-duplex model that can continuously process audio, video and text streams while listening and responding in real time. The model is designed to support interactions in which it can watch, listen and speak simultaneously.
ByteDance has rolled the model out in the Doubao app, moving the technology from a research demonstration toward a consumer-facing product. The launch gives the company another model for real-time multimodal interaction alongside its text and image-generation systems. [Yicai, in Chinese]