CoreReflex today introduces Wind 1.0, a new video generation option built around Seedance 2.0 Standard through Atlas Cloud. Wind brings native 1080p text-to-video and image-to-video generation into CoreReflex, where creators can keep their takes, compare results, and continue editing in the same workspace.
Wind 1.0 is the CoreReflex product name for this integration. The underlying generative model is ByteDance's Seedance 2.0; access is provided through Atlas Cloud.
A new take, with the original still there
A useful creative workflow leaves room for a second attempt. Creators can describe a shot or animate an opening image, review the result, and generate another take without replacing the original.
CoreReflex stores generation history and media references so the brief and its output stay connected. Generated clips appear in the media library and can become part of an editable project alongside narration, music, titles, and real brand assets.
For the launch, we used a trail-running brief to compare a new Wind take with a separately generated second take. The exercise focuses on readable footfalls, consistent clothing and shoes, and a steady tracking camera. Both attempts use the same brief. A second take is a new generation, not an edit that guarantees the first take's identity or motion.
Watch the Wind launch examples.
What is available at launch
The first release supports five-second clips at native 1080p in landscape, portrait, or square format. It accepts text prompts or an opening image, with an optional ending image on the image-to-video route. Generated audio is supported.
Wind is the default for the standard short-video workflow in CoreReflex. Creators can still choose another available model explicitly. The current Wind release does not offer native 4K, source-video editing, or longer clips. Unsupported requests are rejected instead of silently switching to a different model.
A five-second Wind clip currently uses 100 CoreReflex credits. The application shows an estimate before generation; review that estimate for the current price and availability. Each additional take is a separate generation.
Review the footage, then build the film
Native resolution is only one part of a usable shot. Faces, hands, product geometry, motion, and sound still need review. Wind outputs are AI-generated and can contain errors; no prompt guarantees a finished commercial.
Start with one clear action, one camera move, and a short duration. Compare the result against the brief before making more shots. Our guide to writing an AI video prompt explains that approach, and why a video can miss its prompt covers common failure modes.
Try Wind 1.0
Open CoreReflex Studio, select an image, and choose Animate → CoreReflex Wind 1.0. Describe the motion, review the estimate, and generate your first take. Text-to-video is also available through CoreReflex's generation API using the atlas-seedance model identifier.