The system allows users to upload images, video, audio, or text as references for creating videos.

The system allows users to upload images, video, audio, or text as references for creating videos.
The system allows users to upload images, video, audio, or text as references for creating videos.

“Set a visual style using an image, define character movement and camera angle changes with video, create rhythm and atmosphere by uploading a few seconds of audio. Combine all of this with prompts to make the creative process natural, efficient, and truly director-driven,” the company said in its announcement.

Seedance 2.0 is currently available in the Chinese version of the Dreamina service. Access requires a Douyin account (the Chinese version of TikTok). By the end of February, the model is expected to launch on CapCut, Higgsfield, Imagine.Art, and other third-party platforms.

Early users who tested the model praised its strong directing and editing capabilities, highlighting “incredibly natural, almost professional” transitions between scenes.

“There are features that feel almost illegal: you upload a script, and the system generates full scenes—not just clips—with visuals, sound effects, voices, and music. Everything looks polished,” said a user known as el.cine.

The release by ByteDance triggered a rally in Chinese media and AI stocks, according to Bloomberg.
The release by ByteDance triggered a rally in Chinese media and AI stocks, according to Bloomberg.

Shares of COL Group hit the daily trading limit of +20%, while Shanghai Film and Perfect World each rose by 10%.