ByteDance has begun rolling out its new AI video and audio generation model, Dreamina Seedance 2.0, inside its popular editing platform CapCut. Sarah Perez reports for TechCrunch that the model lets creators generate, edit, and sync video and audio content using text prompts, images, or reference videos.
The initial rollout covers Brazil, Indonesia, Malaysia, Mexico, the Philippines, Thailand, and Vietnam. Notably, the United States is not included. ByteDance attributes this to ongoing efforts to address intellectual property concerns, following criticism from Hollywood over alleged copyright infringement.
The model can generate clips of up to 15 seconds across six aspect ratios. It works without reference images and handles realistic textures, movement, and lighting. ByteDance says it performs well with action-focused content, an area where AI video tools have historically struggled.
To address safety concerns, ByteDance has restricted the model from generating video from content that contains real faces. CapCut will also block unauthorized use of copyrighted material. All content produced by the model carries an invisible watermark to help identify AI-generated material shared outside the platform.
Beyond CapCut, the model will also be available on ByteDance’s AI platform Dreamina and its marketing tool Pippit. In China, it is already accessible through the Jianying app.
