While OpenAI appears to be scaling back its video generation ambitions with the closure of its Sora app, ByteDance has announced the rollout of its new audio and video model, Dreamina Seedance 2.0, within its CapCut editing platform. According to the company, this model enables creators to draft, edit, and synchronize video and audio content using simple prompts, images, or reference videos. The release will begin with a phased launch for CapCut users in Brazil, Indonesia, Malaysia, Mexico, the Philippines, Thailand, and Vietnam, with plans to expand to additional markets later.

This launch follows a recent report indicating the model's global rollout was paused to address intellectual property concerns, including criticism from Hollywood over alleged copyright infringement. This likely accounts for the limited initial market availability. In China, the model is already accessible to users of ByteDance's Jianying app.

ByteDance states that the video generation model can operate without any reference images, creating scenes from just a few descriptive words. CapCut excels at rendering realistic textures, movement, and lighting from various visual perspectives, which can be used to edit, enhance, or correct existing footage. The tool also allows creators to experiment with early concepts or sketches before committing to full-scale video production.

Furthermore, Dreamina Seedance 2.0 is designed for a broad spectrum of content, including cooking recipes, fitness tutorials, business overviews, and action-focused videos—areas where AI video models have traditionally struggled. At launch, the model supports video clips up to 15 seconds in length across six different aspect ratios.

Within CapCut, the model will be integrated into various features, including AI Video editing tools and the Video Studio generation suite. It will also be available on ByteDance's AI generation platform, Dreamina, and its marketing platform, Pippit. Due to its capacity for producing highly realistic content, ByteDance has implemented safety measures, preventing the model from generating videos using images or footage containing real faces. CapCut will also restrict the unauthorized generation of copyrighted intellectual property.

All content created with Dreamina Seedance 2.0 will include an invisible watermark to help identify AI-generated material when shared outside the platform. This feature could assist rights holders with takedown requests if copyrighted content is inadvertently produced. ByteDance has committed to collaborating with experts and creative communities to refine and enhance the model's capabilities as it rolls out more widely.