Have you ever imagined editing videos as easily as manipulating text? Well, this idea has now become a reality! ComfyUI-MochiEdit is an open-source video editing tool based on ComfyUI and Genmo Mochi, offering users a novel approach to video editing: converting videos into noise, then resampling the noise using target prompts to generate new videos.
The advantage of this method is that it allows for local editing and video-to-video functionality. This means you can modify just a part of the video without having to process the entire video. Additionally, you can use other videos as references to transform the input video into a new video with specific styles or content.
The core of ComfyUI-MochiEdit consists of two nodes: "Mochi Unsampler" and "Mochi Prepare Sigmas". "Mochi Unsampler" is responsible for converting the input video into noise, while "Mochi Prepare Sigmas" converts the noise back into a video. Users can adjust the parameters of these two nodes, such as noise correction strength, alignment strength with the original video, and the number of steps in the generation process, to control the final video effect.
Furthermore, ComfyUI-MochiEdit supports integration with the ComfyUI-MochiWrapper node, allowing users to adjust the output results through custom prompts for more complex editing needs.
The installation process of ComfyUI-MochiEdit is very simple; users just need to clone it to the specified directory or use the ComfyUI manager for installation, with no additional dependencies required.
The emergence of ComfyUI-MochiEdit brings new possibilities to the field of video editing, making it easy for even those with no video editing experience to create stunning AI videos.
Project link:
https://github.com/logtd/ComfyUI-MochiEdit?tab=readme-ov-file#mochi-unsampler
