Topic

OpenStoryline: The Next-Generation Paradigm for Video Editing

Over the past decade, video editing tools have significantly lowered the barrier to entry, evolving from professional software to platforms accessible to everyday creators. Tools such as Jianying (CapCut) enable ordinary users to complete basic editing tasks, yet the creative workflow still revolves around complex timeline operations. For many UGC and lightweight PGC creators, video editing remains time-consuming and relatively demanding. At the same time, real-world editing workflows contain many steps that can be structured, such as highlight selection, script generation, rhythm control, and template-based editing. These repetitive tasks create strong opportunities for AI-driven automation. Against this background, we began exploring a new paradigm: enabling users to interact with the system through natural language, allowing AI to complete the entire process from content understanding to video generation. This talk shares our practical journey from real business workflows to the development of the video editing agent OpenStoryline. We will discuss key challenges in multi-turn dialogue, task planning, and tool execution, and explore why “editing agents” may become one of the next important directions for AI-powered video creation. Outline: 1. The current state of video editing: has the low-barrier era truly arrived? 2. Why we need even lower barriers to video editing 3. Teaching AI to edit videos: our experience with automated editing 4. From workflow to agent: the birth of a video editing agent 5. Why intelligent editing agents may become the next major AI trend