One Button Transformation
Create transformation videos where a cube, sphere, capsule, or custom shape appears on a table with a button on top. A hand presses the button and the object transforms into a person, character, or item based on your reference images. The workflow automatically uses the colors and theme of your subject to generate matching first and last frames while keeping the environment and style consistent throughout the transformation. Simply provide your reference images, choose a shape, generate the frames, and feed them into Seedance to create the final animation effect. Includes support for both Seedance 2.0 for higher-quality results and Seedance 1.5 as a more affordable alternative that still produces solid results.
简介
One Button Transformation is a guided ComfyUI pipeline that turns a simple tabletop prop into your subject with a single press, then animates the metamorphosis. You provide one or more reference images of a person, character, or product; the workflow builds a color-matched first frame where your chosen shape (cube, sphere, capsule, or a custom shape you describe) sits on a table with a button. It then creates a stylistically consistent last frame where your subject appears in the same scene, ready for a seamless transformation.
你需要提供
- 正在读取工作流
在 ComfyUI 中运行。 此工作流无法在本页面运行。你可以在 Comfy Cloud 中打开（已预先加载所需内容），或下载后在自己的 ComfyUI 中使用。
在 Cloud 中试用常见问答
Describe the shape directly in the prompt within the First Frame group (e.g., “a matte white capsule with a small black button on top, on a wooden table”). The GeminiNode/GeminiNanoBanana2V2 block uses this text to render the prop. Use terms like cube, sphere, capsule, or a custom shape description.
Both take your first and last frames and synthesize in-betweens. The Seedance 2.0 group (ByteDance2FirstLastFrameNode) generally produces higher-quality motion and detail, while the Seedance 1.5 Version (ByteDanceFirstLastFrameNode) is faster and more budget-friendly. Start with 2.0 for hero shots and 1.5 for quick iterations.
Use the same subject reference(s) for both steps, avoid changing the scene description between the First Frame and Last Frame groups, and keep camera and lighting phrasing stable. If available, reuse the same seed. The workflow is designed to preserve palette and scene continuity from your inputs.
Yes. Describe the custom shape clearly in your prompt (e.g., “low-profile hexagonal puck”) and include material and color cues. If you have a specific silhouette image, load it via LoadImage as a guiding reference where the workflow allows.