Cassette-Editor：一款像人一样剪辑的视频编辑 Agent
Cassette-Editor: A Video Editing Agent That Edits the Way You Do
AI 导读
arXiv 论文提出 Cassette-Editor，一个仅含一读一写两个工具接口的视频编辑智能体。所有写入都经过与用户手动编辑相同的执行器、规则与提交服务，要么原子提交，要么返回定位诊断和修复选项。在 277 个被破坏的自有提交编辑的实验中，模型几乎能全部修复，定位问题使修复所需 input tokens 减至不到一半。
正文 · 原文
Abstract:Language agents increasingly work by calling tools on a persistent artifact, such as a code repository, and the result of each call shapes their next step. Video editing is a demanding case of this pattern: what a frame shows depends on how every clip is trimmed, retimed, cropped, and nested, so knowledge about the footage holds only for particular uses of it, and an invalid edit can leave the timeline structurally broken. Existing video editing agents keep their edits valid with critic agents, reviewers, or restricted output formats. We present Cassette-Editor, a video editing agent with a deliberately simple interface: one tool reads the project and one tool writes to it. Every write passes through the same executor, rules, and commit service as the user's manual edits, and either commits atomically or is rejected with located diagnostics and repair options, while the model keeps the creative choices and decides how to recover. Reads relate stored analysis of the footage to each use of it on the timeline. In a controlled experiment, a real model is shown 277 of its own committed edits after damaging them. It corrected nearly every call with full or minimal feedback, but locating each problem cut the input tokens a correction needed to less than half, and its corrections dropped other values the model had written in one case rather than seven.
| Comments: | 8 pages, 4 figures, 2 tables. Preprint; evaluation to be extended in a future version |
| Subjects: | Multimedia (cs.MM); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.37690 [cs.MM] |
| (or arXiv:2610.37690v1 [cs.MM] for this version) | |
| https://doi.org/10.48550/arXiv.2610.37690 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Zhixin Zeng [view email]
[v1]
Mon, 19 Oct 2026 14:27:55 UTC (415 KB)
来源：arXiv cs.AI · arxiv.org