LongCatAvatar_Z-Image_DynmicLength(Loops) Ver. 20251223 (Workflow) (Patreon Supporters) (Patreon)

LongCat-Video-Avatar

https://meigen-ai.github.io/LongCat-Video-Avatar/

For ComfyUI AI Models:

Diffusion Model:

LongCat-Avatar_comfy_bf16.safetensors

https://huggingface.co/Kijai/LongCat-Video_comfy/blob/main/Avatar/LongCat-Avatar_comfy_bf16.safetensors

LoRA:

LongCat_distill_lora_alpha64_bf16.safetensors

https://huggingface.co/Kijai/LongCat-Video_comfy/blob/main/LongCat_distill_lora_alpha64_bf16.safetensors

Text Encoder:

umt5-xxl-enc-bf16.safetensors

https://huggingface.co/Kijai/WanVideo_comfy/blob/main/umt5-xxl-enc-bf16.safetensors

Or

umt5_xxl

https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/tree/main/split_files/text_encoders

VAE:

wan_2.1_vae.safetensors

https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/blob/main/split_files/vae/wan_2.1_vae.safetensors

wav2vec2 Use Auto Download

Or

save in ComfyUI/models/wav2vec2

https://huggingface.co/Kijai/wav2vec2_safetensors/tree/main

MelBandRoFormer

https://huggingface.co/Kijai/MelBandRoFormer_comfy/tree/main

save in ComfyUI/models/diffusion_models

For ComfyUI Custom Node :

Wrapper node (longcat_avatar branch)

https://github.com/kijai/ComfyUI-WanVideoWrapper/tree/longcat_avatar

ComfyUI-MelBandRoFormer

https://github.com/kijai/ComfyUI-MelBandRoFormer

Video Basic Usage for LongCat Avatar : https://youtu.be/eZTdaLbzqL4

Workflow : LongCatAvatar_Z-Image_DynmicLength(Loops) Ver. 20251223

Step to start:

Use Z-Image Turbo create your avatar image if you want. (optional)

Make sure you have Z-Image Turbo (check here if you don't know what is it.)

Here you have an option for Bypass Z-Image Turbo groups if you are not using it.

Basic video setting , and text prompt for Z-Image and Video text prompt. Again if you have your own image already, turn Z-Image Turbo Off using the group Bypasser option.

If you use Z-Image Turbo, after 2nd pass sampling, you are going have your image. Generate until you like it and Turn On "LongCat Avatar Input Image" group, and Bypass the Load Image node, because you are passing Z-Image's generated image to the avatar generation.

Otherwise, use your own image by using Load Image node , ignore this Z-Image Turbo step.

Audio setup.

2 options, use TTS, I use ChatterBox TTS in this case. You can try other TTS if you like.

or...

Load Audio file..

Bypass either one that are not using.

Once you select right model files, input correct files, text and setting, you will able to generate something like this:

Have Fun :)