
China’s Homegrown L4 Intelligent Agent Platform MasterAgent Fully Opens to the Public
Want to read in a language you're more familiar with?
MasterAgent marks a major step in China’s AI autonomy push, bringing a fully domestic, L4-level multi-agent platform into real-world, large-scale use.
On December 30, 2025, MasterAgent, an L4-level intelligent agent master system developed by Shenyuan AI, was officially made fully accessible to users.
Built entirely on self-developed, full-stack technologies, MasterAgent establishes a comprehensive domestic technology chain, spanning chip adaptation, core infrastructure, and system architecture. By eliminating reliance on overseas technologies, the platform meets strict data security and compliance requirements in sensitive sectors such as government and finance, filling a long-standing gap in China’s high-end multi-agent collaboration platforms.

MasterAgent features a framework-based design that enables users to create multi-agent clusters within minutes using natural language commands. The system has already been deployed across multiple industries: it has powered marketing agents for Boway Group, enabled the “Ligeyi Legal AI Agent” for the legal sector, and—during its limited beta phase—served hundreds of government and enterprise clients, including Haier and GAC. With full public access now available, individual users and SMEs can connect to the platform and build customized intelligent agents.

The platform’s capabilities have been validated through high-profile use cases. MasterAgent was selected as the primary AI Agent technology service provider for the 15th National Games opening ceremony, where it generated a 174-meter-long AI-powered digital scroll. Shenyuan AI has also recently been recognized on several industry rankings, including “China’s Top 20 AI Agent Companies for Commercial Deployment.”
Looking ahead, MasterAgent aims to accelerate the large-scale adoption of domestic intelligent agents, driving China’s AI industry beyond simple task execution toward end-to-end outcome delivery.




