Kosinkadink /ComfyUI-VideoHelperSuite

About ComfyUI-VideoHelperSuite

Load a video or an image folder as frames, then write those frames back out as video. These nodes live in ComfyUI, a node-based app for running AI models. You get steady frame handling and encoding for AI video graphs. Encoding needs a working ffmpeg build, and the docs assume you already know ComfyUI.

Best for: Pipeline builders who need dependable video and image-sequence loading, batching and encoding inside AI video graphs.

• Loads a video into an image batch with frame rate, size, cap and skip controls
• Reads videos or image sequences from outside paths with the path node variants
• Writes image batches to video with frame rate, loop count, ping-pong and format
• Mixes an audio file into the output video when you connect the audio input
• Splits, merges, duplicates and counts batches with latent and image utility nodes
• Adds your own output formats as JSON files in the video_formats folder

1. 1
   Add a Load Video node and point it at your video file.
2. 2
   Set force_rate, force_size, frame_load_cap and skip_first_frames to pick the frames you want.
3. 3
   Send the image batch through your generation nodes, then into a Video Combine node.
4. 4
   In Video Combine set frame_rate, pick a format, and connect an audio input if you want sound in the file.
5. 5
   Turn on VHS Advanced Previews from the settings gear next to Queue Prompt to see previews that match your node settings.