Supporters of Marcus Endicott’s Patreon can access weekly or monthly video consultations on this topic.
Quwan Technology (Guangzhou Quwan Network Technology Co., Ltd., 广州趣丸网络科技有限公司, 趣丸科技) is a Guangzhou-based technology company whose initial team was formed on September 23, 2013 and whose formal establishment date is December 13, 2014. On its official site, the company presents itself as an innovative technology enterprise spanning interest-based social networking, esports, and artificial intelligence, with founder Song Ke serving as CEO. That background is important in the digital human context because Quwan did not emerge as a narrow virtual character studio or a single-product digital human startup. It developed first through larger-scale consumer internet businesses and then extended into AI, which places it in the category of Chinese platform companies that expanded from internet products into applied AI infrastructure and commercially deployable AI systems.
In the digital human field, Quwan should be understood as a broader applied AI company that has made digital humans part of a larger commercial technology stack through its AI arm, Quwan Smart (趣丸智能). The company has described this work as an enterprise-oriented AI interaction business and has said it built a multimodal interactive digital human generation platform in collaboration with the Chinese University of Hong Kong, Shenzhen and HKUST (Guangzhou). Public descriptions of its system state that the platform integrates self-developed high-precision 3D deformation algorithms, an automated “algorithm + rendering + driving” workflow, and multimodal network models used for scene construction, avatar generation, and digital human livestreaming. Quwan has also publicly claimed that its system can generate highly natural digital humans in roughly 10 seconds from one or several photos, with facial similarity above 90 percent, while supporting multimodal interaction across expressions, text, voice, and body movement.
The company positions these digital humans as practical commercial infrastructure rather than purely promotional or entertainment figures. Reported application scenarios include science education, livestream retail, animation and games, cultural and museum settings, smart-city use cases, and telecom service operations. One named example is the China Telecom service-hall digital human Xiaoxin (小信), which has been described as supporting real-time voice interaction, guided consultation, and one-stop service handling in service hall environments. Quwan’s broader AI stack is also tied to its speech technology work, including the MaskGCT model released with CUHK-Shenzhen in October 2024 for voice cloning, multilingual synthesis, and controllable speech generation. In commercial terms, a 2025 Guangdong media report linked Quwan’s digital human and AI business to e-commerce livestream operations with cumulative client GMV exceeding RMB 1 billion. Taken together, Quwan’s role in digital humans is best described as that of a scaled AI applications company using digital humans as one of its main interface layers for enterprise service, livestream commerce, and multimodal human-computer interaction.