Reproduce “CoworkOS:Scaling Agent Coworkers for Real-world Automation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jiajie Jin, Guanting Dong, Chengjun Pan, Xin Deng, Zhuohui Sheng, Guoqiang Wang, Tao Gui, Yutao Zhu and 32 more
CoworkOS addresses complex work by letting a Main Agent direct delegated tasks while a Coordinator tracks assignments and feedback, and Subagents execute bounded work. Persistent jobs, memory, and shared resources support continuity across interruptions. In a five-round post-training case study, the mean across eight benchmark scores rose from 33.19 to 42.95; this measured increase does not isolate CoworkOS’s causal effect.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.