ClawTeam by the Data Intelligence Lab at the University of Hong Kong (HKUDS) is an autonomous multi-agent orchestration framework built on the principle of “One Command → Full Automation.” A leader agent decomposes a high-level goal into subtasks and orchestrates a team of specialized sub-agents to execute them in parallel, with humans providing only the initial objective.
At scale, ClawTeam can coordinate 8 specialized sub-agents across 8 NVIDIA H100 GPUs, making it one of the most computationally ambitious open-source agent orchestration systems available. This hardware-parallel design is targeted at research automation workflows where different agents might handle literature review, data analysis, code execution, and result synthesis simultaneously.
ClawTeam serves as the upstream for win4r’s ClawTeam-OpenClaw fork, which adds production hardening and deep OpenClaw integration. The HKUDS project remains focused on research-grade automation and academic benchmarking, while community forks like win4r’s adapt it for operational deployment. Together they represent one of the most sophisticated multi-agent architectures in the OpenClaw ecosystem.