Project Page: <a href=\"https://wucy0519.github.io/MMLVE/\" rel=\"nofollow\">https://wucy0519.github.io/MMLVE/</a><br>Source Codes: <a href=\"https://github.com/Wucy0519/MMLVE\" rel=\"nofollow\">https://github.com/Wucy0519/MMLVE</a><br>Benchmark: <a href=\"https://huggingface.co/datasets/wcy1234567/MMLVE-Bench\">https://huggingface.co/datasets/wcy1234567/MMLVE-Bench</a></p>\n","updatedAt":"2026-08-28T04:26:31.659Z","author":{"_id":"6449f2dfeb7db8f70fb990f8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6449f2dfeb7db8f70fb990f8/limS6x8txJJHCIBDbih-N.png","fullname":"Fuchen","name":"FireCRT","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":1,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.6021734476089478},"editors":["FireCRT"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/6449f2dfeb7db8f70fb990f8/limS6x8txJJHCIBDbih-N.png"],"reactions":[{"reaction":"👍","users":["kira7x"],"count":1}],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.26809","authors":[{"_id":"6a910d24a64059bab69c36ba","name":"Chenyang Wu","hidden":false},{"_id":"6a910d24a64059bab69c36bb","name":"Fuchen Long","hidden":false},{"_id":"6a910d24a64059bab69c36bc","name":"Binyuan Huang","hidden":false},{"_id":"6a910d24a64059bab69c36bd","name":"Xinlong Sun","hidden":false},{"_id":"6a910d24a64059bab69c36be","name":"Xi Chen","hidden":false},{"_id":"6a910d24a64059bab69c36bf","name":"Chun-Le Guo","hidden":false},{"_id":"6a910d24a64059bab69c36c0","name":"Chongyi Li","hidden":false}],"publishedAt":"2026-08-27T00:00:00.000Z","submittedOnDailyAt":"2026-08-28T00:00:00.000Z","title":"Thinking on Shots: Consistent Multi-Shot Video Editing with Agentic Reasoning","submittedOnDailyBy":{"_id":"6449f2dfeb7db8f70fb990f8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6449f2dfeb7db8f70fb990f8/limS6x8txJJHCIBDbih-N.png","isPro":false,"fullname":"Fuchen","user":"FireCRT","type":"user","name":"FireCRT"},"summary":"While generative AI has significantly advanced video editing, existing methods primarily focus on single-shot or short video clips. Editing long videos with multiple instructions remains a formidable challenge. Naive chunking strategies, e.g., fixed-duration segmentation, often lead to entity fragmentation, severe editing hallucinations, and disrupted temporal continuity. To bridge this gap, we introduce the Multi-Instruction Multi-Shot Long-Video Editing (MMLVE) task, which is structured around three core objectives: Cross-Shot Editing Consistency (CSEC), Multi-Instruction Decoupling (MID), and Zero-Destruction on Spatiotemporal Structure (ZDSS). To tackle these three unique challenges, we introduce an agentic editing framework that leverages the synergy of Large Language Models (LLMs) and Vision-Language Models (VLMs) to achieve shot-level video decoupling and precise instruction parsing. Furthermore, to comprehensively evaluate this task, we construct MMLVE-Bench, which is an MMLVE-focused dataset characterized by complex real-world spatiotemporal dynamics, high-density heterogeneous instructions, and sparse, random entity distributions. Three MMLVE-focused evaluation metrics are further exploited to assess the quality of the editing results. Extensive experiments demonstrate that our MMLVE-Agent outperforms existing closed-source SOTA approaches (e.g., Seedance 2.0), successfully eliminating editing hallucinations, preserving cross-shot editing consistency, and attaining seamless spatiotemporal transitions.","upvotes":3,"discussionId":"6a910d24a64059bab69c36c1","projectPage":"https://wucy0519.github.io/MMLVE/","githubRepo":"https://github.com/Wucy0519/MMLVE","githubRepoAddedBy":"user","ai_summary":"An agentic framework combining LLMs and VLMs enables consistent, multi-instruction editing of long multi-shot videos while preserving spatiotemporal structure.","ai_keywords":["Multi-Instruction Multi-Shot Long-Video Editing","Cross-Shot Editing Consistency","Multi-Instruction Decoupling","Zero-Destruction on Spatiotemporal Structure","Large Language Models","Vision-Language Models","MMLVE-Bench"],"ai_summary_model":"thinkingmachines/Inkling-Small","githubStars":2,"organization":{"_id":"661ceb12e7b0ab12bc5f8b8d","name":"NankaiUniversity","fullname":"Nankai University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/f6J0Jh9AWN1FQja8TYosr.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"6449f2dfeb7db8f70fb990f8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6449f2dfeb7db8f70fb990f8/limS6x8txJJHCIBDbih-N.png","isPro":false,"fullname":"Fuchen","user":"FireCRT","type":"user"},{"_id":"6878987600ce33ca49b91d3f","avatarUrl":"/avatars/e30e08c2f8bfdb4f00103327dd43c21a.svg","isPro":false,"fullname":"cy","user":"wuucy","type":"user"},{"_id":"6a27b06ce75f45d33e8c4cd1","avatarUrl":"/avatars/951f6dceeec9e31ae9ebdeab128fd96e.svg","isPro":false,"fullname":"chenggh","user":"chenggh","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"661ceb12e7b0ab12bc5f8b8d","name":"NankaiUniversity","fullname":"Nankai University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/f6J0Jh9AWN1FQja8TYosr.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.26809.md","query":{}}">
Thinking on Shots: Consistent Multi-Shot Video Editing with Agentic Reasoning
Published on Aug 27
· Submitted by Fuchen on Aug 28 Abstract
An agentic framework combining LLMs and VLMs enables consistent, multi-instruction editing of long multi-shot videos while preserving spatiotemporal structure.
While generative AI has significantly advanced video editing, existing methods primarily focus on single-shot or short video clips. Editing long videos with multiple instructions remains a formidable challenge. Naive chunking strategies, e.g., fixed-duration segmentation, often lead to entity fragmentation, severe editing hallucinations, and disrupted temporal continuity. To bridge this gap, we introduce the Multi-Instruction Multi-Shot Long-Video Editing (MMLVE) task, which is structured around three core objectives: Cross-Shot Editing Consistency (CSEC), Multi-Instruction Decoupling (MID), and Zero-Destruction on Spatiotemporal Structure (ZDSS). To tackle these three unique challenges, we introduce an agentic editing framework that leverages the synergy of Large Language Models (LLMs) and Vision-Language Models (VLMs) to achieve shot-level video decoupling and precise instruction parsing. Furthermore, to comprehensively evaluate this task, we construct MMLVE-Bench, which is an MMLVE-focused dataset characterized by complex real-world spatiotemporal dynamics, high-density heterogeneous instructions, and sparse, random entity distributions. Three MMLVE-focused evaluation metrics are further exploited to assess the quality of the editing results. Extensive experiments demonstrate that our MMLVE-Agent outperforms existing closed-source SOTA approaches (e.g., Seedance 2.0), successfully eliminating editing hallucinations, preserving cross-shot editing consistency, and attaining seamless spatiotemporal transitions.
Community
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2608.26809 in a model README.md to link it from this page.
Cite arxiv.org/abs/2608.26809 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2608.26809 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.