이미지: ElevenLabs (YouTube) 영상 갈무리 · METAL LAB 편집
Summary
- ElevenLabs has added ByteDance Seed's video generation model Seedance 2.5 to its creative tool ElevenCreative
- ElevenLabs says the same interface can now chain together voiceover, lip-sync, music, and sound effects
ElevenLabs has announced that it's added ByteDance Seed's video generation model, Seedance 2.5, to ElevenCreative, its video and image production tool. In a demo video posted to YouTube, ElevenLabs explained that the model can take in up to 50 reference images at once, hold character, product, and lighting details steady throughout, and generate a full 30-second clip in a single pass — no cuts, no stitching.
That framing takes direct aim at a known weak spot in AI video generation. Longer AI-generated videos have typically been built by stitching together several short clips, and that process often introduces small inconsistencies — a face that shifts slightly between cuts, a product shape that drifts from one scene to the next. ElevenLabs itself put it plainly in the video: the model "forgets what it was making," a reference to how context breaks down when short generations are pieced together one after another. Seedance 2.5 takes a different approach, feeding the model reference images for characters, products, locations, and lighting up front, then describing the whole scene in a single prompt so consistency holds within one continuous generation rather than across several stitched ones.
Seedance is a family of video generation models built by ByteDance Seed, the AI research arm of TikTok's parent company ByteDance, and ElevenLabs is integrating that model into its own tool rather than building the underlying video model itself. Around the same time, other video production tools like Higgsfield and Krea have been rolling out their own integrations of Seedance 2.5, pairing it with camera-path and 3D scene controls. What sets ElevenLabs' announcement apart is that it bundles audio post-production into the same workflow.
According to ElevenLabs, once a video is generated inside ElevenCreative, users can immediately add voiceover, lip-sync, background music, and sound effects without leaving the tool. That means creators no longer have to bounce between separate apps for video generation and audio post-production. That said, finer details like actual output quality or how accurately the model honors reference images aren't something this demo video alone can confirm.
Comments