Seedance 2.0 released

Higgsfield has put Seedance 2.0 on its platform, turning a previously limited video model into a public product push aimed at filmmakers and creators. (higgsfield.ai) Seedance 2.0 is a text-to-video system that can also take images, video clips and audio as references, then generate multi-shot video with synced sound. Higgsfield’s product page says a single generation can use up to 12 assets and keep characters consistent across cuts. (higgsfield.ai) The model itself comes from ByteDance’s Seed team, and ByteDance describes it as a multimodal audio-video generator built to follow visual composition, camera movement, motion rhythm and sound cues from mixed inputs. Higgsfield is packaging that model inside its own creator workflow and publishing tutorials around it. (seed.bytedance.com) (higgsfield.ai) Higgsfield’s launch pitch centers on volume and control rather than one-off clips. Its site advertises “one-click video recreation,” frame-level editing, native audio sync and a “Cinema Studio” workflow for assembling cinematic scenes faster. (higgsfield.ai 1) (higgsfield.ai 2) That framing comes after months of competition among artificial intelligence video companies to move from short demo shots to usable production tools. Rivals including Runway, Google and Kling have all pushed video models with stronger editing, reference handling and longer-form storytelling claims, while Higgsfield has leaned into presets and filmmaking interfaces on its YouTube channel and product pages. (youtube.com) (higgsfield.ai) Higgsfield tied the release to a showcase project called “Zephyr,” which it posted on YouTube as a full episode and separately as a short intro. The company says the project was generated inside Higgsfield with Seedance 2.0 and presents it as the start of an “Original Series” strategy rather than a single tech demo. (youtube.com 1) (youtube.com 2) In a how-to post, Higgsfield said “Zephyr” was built as a sci-fi story with five original characters, a monster invasion and an original K-pop sequence, and it published prompts and workflow steps for users who want to recreate the process. That turns the release into both a model launch and a template for serialized synthetic entertainment. (higgsfield.ai) Access has shifted quickly. Higgsfield said in an April 2 post that Seedance 2.0 was available through business email verification outside the United States and Japan, but the newer product pages and the “Zephyr” video now say unlimited Seedance 2.0 access is live more broadly. (higgsfield.ai) (youtube.com) The company and affiliated promotional pages have also pushed limited-time discounts for new users, with claims ranging from 65% to 70% off depending on the offer page. Higgsfield’s own materials visible in the “Zephyr” release describe “up to 70% OFF” for a limited time. (youtube.com) (sci-tech-today.com) The release does not settle the larger questions around artificial intelligence video, including copyright, disclosure and how much of a finished episode can be made reliably without heavy cleanup. But Higgsfield is clearly betting that the next sales pitch is not a prettier clip; it is a repeatable pipeline for making many of them. (seed.bytedance.com) (higgsfield.ai)