HunyuanVideo15SuperResolution - ComfyUI Built-in Node Documentation

Complete documentation for the HunyuanVideo15SuperResolution node in ComfyUI. Learn its inputs, outputs, parameters and usage.

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

The HunyuanVideo15SuperResolution node prepares conditioning data for a video super-resolution process. It takes a latent representation of a video and, optionally, a starting image, and packages them along with noise augmentation and CLIP vision data into a format that can be used by a model to generate a higher-resolution output.

Inputs

Parameter	Data Type	Required	Range	Description
positive	CONDITIONING	Yes	N/A	The positive conditioning input to be modified with latent and augmentation data.
negative	CONDITIONING	Yes	N/A	The negative conditioning input to be modified with latent and augmentation data.
vae	VAE	No	N/A	The VAE used to encode the optional start_image. Required if start_image is provided.
start_image	IMAGE	No	N/A	An optional starting image to guide the super-resolution. If provided, it will be upscaled and encoded into the conditioning latent.
clip_vision_output	CLIP_VISION_OUTPUT	No	N/A	Optional CLIP vision embeddings to add to the conditioning.
latent	LATENT	Yes	N/A	The input latent video representation that will be incorporated into the conditioning.
noise_augmentation	FLOAT	No	0.0 - 1.0	The strength of noise augmentation to apply to the conditioning (default: 0.70).

Note: If you provide a start_image, you must also connect a vae for it to be encoded. The start_image will be automatically upscaled to match the dimensions implied by the input latent.

Outputs

Output Name	Data Type	Description
positive	CONDITIONING	The modified positive conditioning, now containing the concatenated latent, noise augmentation, and optional CLIP vision data.
negative	CONDITIONING	The modified negative conditioning, now containing the concatenated latent, noise augmentation, and optional CLIP vision data.
latent	LATENT	The input latent is passed through unchanged.