VIVID:让 Codex、Claude Code 等编程智能体直接剪辑视频的视觉工具框架
VIVID: A Visual Harness for Agentic Video Editing
用通用编程智能体剪视频这件事有人真做出来了。VIVID 是个外挂框架,Codex 加上它分数涨 20 多个百分点,还能配 Claude Code 和 OpenCode 用。做视频工具或智能体工作流的开发者可以研究下这套观察-剪辑-检视的闭环设计。
VIVID 是一个无需额外模型训练的视觉 harness,可挂载到原生编程智能体上完成视频后期制作。在包含 Assembly、Repair、Sequencing、Repurpose 四类共 100 个任务的 AgenticVBench 上,Codex 搭配 VIVID 平均得分 58.6%,比原生 Codex 提升 20.2 个百分点。
技术上,VIVID 提供按需视频检视、音频分析和转写工具,并把每条观察记录与其对应的媒体版本和时间坐标一起保存。智能体在推理过程中可主动检索这些观察,用自带的编码和 shell 工具执行剪辑,再检视渲染结果来指导下轮修改,形成完整的反馈闭环。
同一 harness 还兼容 Claude Code 和 OpenCode,平均得分分别为 55.1% 和 37.1%,均大幅超过各自原生基线。这说明视频剪辑能力可以作为一种可复用的外挂框架,而非绑定某个特定模型。
VIVID: A Visual Harness for Agentic Video Editing
Video editing requires agents to make editing decisions grounded in audiovisual observations and precise temporal localization, and to evaluate their effects in rendered outputs. General-purpose coding agents possess strong comprehension, reasoning, and code execution capabilities, while media interaction and evidence management mechanisms tailored to video tasks can further unlock their editing potential. We introduce VIVID, a visual harness that supports multiple native coding agents in video post-production without additional model training. VIVID provides tools for on-demand video inspection, audio analysis, and transcription, and preserves observations alongside their corresponding media versions and temporal coordinates. Agents can actively retrieve these observations as they reason, perform edits using their native coding and shell tools, and inspect rendered outputs to guide subsequent revisions. On AgenticVBench, comprising 100 tasks across Assembly, Repair, Sequencing, and Repurpose, Codex with VIVID achieves an average score of 58.6%, an improvement of 20.2 percentage points over native Codex. The same harness also integrates with Claude Code and OpenCode, achieving average scores of 55.1% and 37.1%, respectively, both substantially exceeding their native baselines. These results highlight VIVID's potential as a reusable visual harness for extending the capabilities of general-purpose coding agents to video editing.
This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include VIVID, Codex, Claude Code, OpenCode.
Source attribution: curated from arXiv cs.AI (@Wenhan Yu, Zhaoxi Zhang, Zhilong Tan, Weikang Li, Lei Sha). Cross-source corroboration: 2 independent sources have covered this development (including 小胡子哥), confirming industry-wide signal strength.
- 小胡子哥11-15 12:51Source