AI 破框特效示例视频

关于这个示例

上方视频是AI 破框特效的示例。AI 破框特效：需要一张源图片和预设提示词，所需素材一目了然。默认使用 Seedance 2 Fast，支持 16:9 / 9:16 / 1:1 画面比例和5秒 / 8秒时长。

制作类似的视频

上传 1 张主体清晰的照片，即可制作你自己的版本。你也可以额外添加 1 张参考图（可选）。视频时长为 5 秒 / 8 秒，画面比例为 16:9 / 9:16 / 1:1。每次生成 75 积分起，最终消耗取决于你选择的模型、时长和设置。

提示词示例

AI 破框特效从这段提示词开始。你可以直接使用，也可以在生成前加入自己的要求。

Use the input image as the FIRST FRAME and keep it consistent. LOCK the camera completely (no zoom, no pan, no tilt). Keep the background, lighting, and environment perfectly static across the entire video. Replace the frame rule with an AUTO-FIT FRAME rule: the picture frame MUST always fully contain the entire subject at the start, no matter how large or tall the subject is. If needed, the frame automatically expands or adjusts its aspect ratio to fit the whole subject inside WITHOUT cropping. The frame MUST remain a rigid, perfectly rectangular frame (no bending, no warping, no melting), fixed on the wall, same material and style throughout. The frame can only resize ONCE at the very beginning (0-10%) to fit the subject, then stays completely unchanged afterward. The subject MUST remain exactly the same as the input (same identity, texture, colors, proportions). No stylization. No redesign. No duplicates. Animation: the subject slowly climbs OUT of the frame while the frame stays on the wall. 0-25%: the subject leans forward and places ONE limb/edge over the bottom frame border, clearly in front of the frame. 25-55%: a second contact point crosses the frame edge; the torso/upper…

生成这个特效