Seedance 2.0 officially opens application for general API customers

ByteDance's Volcano Engine launched the open application for Seedance 2.0 to general API customers on April 2nd, marking the official expansion of this multimodal video generation model from the Volcano Fangzhou Experience Center and limited invitation testing phase to a broader audience of developers and enterprises.

Seedance 2.0 was launched in the Volcano Fangzhou Experience Center on February 12, 2026. It uses a multimodal audio-video joint generation architecture and supports four modalities: text, images, audio, and video as input. It can achieve highly consistent character retention, director-level camera control, realistic physical simulation, and native audio-visual synchronization. Users can combine multiple images, video clips, and audio reference materials to generate video segments with a cinematic feel.

Compared to earlier versions, Seedance 2.0 has significantly improved motion replication, character consistency, and audio-visual stability, and is considered one of the most controllable AI video models in the industry. It supports text-to-video, image-to-video, and multimodal hybrid references, and is suitable for short drama production, e-commerce marketing videos, special effects templates, and professional film and television content creation. Volcano Engine emphasized its "production-grade" features, making it suitable for complex creative needs that require precise control.

Previously, Seedance 2.0 mainly offered free or limited experience, and the API service was originally planned to launch in mid-to-late February, but was delayed due to factors such as copyright compliance and content security. Starting April 2nd, developers can submit API open application through the Volcano Fangzhou console, mainly targeting entities that have completed enterprise certification.

On the same day, the third-party platform Invideo announced the launch of Seedance 2.0 for Max, Generative, and Team plan users, supporting commercial email verification (excluding some areas in the U.S. and Japan). Other integrated platforms also opened access, further lowering the usage threshold. This shows that Volcano Engine is accelerating its transition from a closed experience to an open ecosystem.

The technical highlights lie in the multimodal joint architecture and strong reference consistency: precise camera movement, physical simulation, coherent multi-character scenes, reducing issues such as flickering or drifting commonly seen in traditional AI videos. With the API opening, Seedance 2.0 is expected to drive AI video from entertainment experiments to commercial productivity tools, especially unlocking potential in short videos, e-commerce, and education.

Copyright protection remains a key focus in the industry. The current launch is expected to include strict content review mechanisms. Developers can submit applications by logging into the Volcano Engine official website or the Volcano Fangzhou platform, and pay attention to the official documentation and pricing. This release marks an important move by ByteDance in the generative AI video field, building a more practical video generation infrastructure.