ComfyUI-VideoHelperSuite

Why we included this project

ComfyUI users who work with video will recognize this as the standard set of nodes for moving footage in and out of a graph. The Load Video node converts a clip into individual frames you can hand to image or latent models, and Video Combine reassembles the generated frames into a playable clip, with optional audio and loop or ping-pong playback. The settings are built around animation work: you can force a target frame rate to line up with tools like AnimateDiff, cap the batch size, skip leading frames to process a long input in chunks, and resize while keeping the aspect ratio. There are also utilities for splitting, merging, duplicating, and counting latents or images, plus animated previews that reflect your current node settings so you can see exactly which slice of a source video you selected. If you are building text-to-video or frame-interpolation pipelines, these nodes are the practical bridge between your source material and your model outputs.