Reproduce “Auto-Director: Leveraging Multi-Agent Collaboration for Autonomous Video Editing” 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
Yuxia Chen, Wenbo Zhu, Yangguang Ji, Zhenxiang Jiang, Teng Zhong, Yu Zhang
Auto-Director coordinates factual planning, footage retrieval, editing, and revision to make narrated videos from open-ended topics. Its agents check claims, organize footage into searchable shots, select and reframe clips, and audit scenes to route targeted corrections. On VEdit-Bench, it scores 4.34 for factual accuracy versus VideoAgent’s 3.34, leading automated systems on five dimensions; this evidence is limited to the benchmark.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.