The-frizzy1/Wan22ANIMATE

Wan2.2 Animate GGUF — Unlimited Video Length + Looping

By: The_frizzy1 Hardware target: RTX 3050 (4 GB) / RTX 3060 (12 GB) CivitAI: https://civitai.com/models/2046477/wan22-animate-gguf-workflow-unlimited-video-length-looping-mechanism YouTube: https://www.youtube.com/@the_frizzy1

🎥 Video Explainer: https://www.youtube.com/watch?v=rtyfdmL-wF4

---

What This Is

Wan2.2 Animate 14B in GGUF format with a looping mechanism for unlimited-length video generation.

• 12 GB VRAM: up to Q8, 25 frames per loop
• 4 GB VRAM: up to Q4, 25 frames per loop

---

What's New in v1.2

• Looping mechanism for unlimited video length
• Points editor out of subgraph
• Better organisation

---

Required Custom Nodes

Node	Link
comfyui_controlnet_aux	https://github.com/Fannovel16/comfyui_controlnet_aux
KJNodes	https://github.com/kijai/ComfyUI-KJNodes
segment-anything-2	https://github.com/kijai/ComfyUI-segment-anything-2
IAMCCS-nodes (v1 only)	https://github.com/IAMCCS/IAMCCS-nodes
VideoHelperSuite	https://github.com/Kosinkadink/ComfyUI-VideoHelperSuite
Execution Inversion Demo	https://github.com/BadCafeCode/execution-inversion-demo-comfyui

---

Model Downloads

Main model: https://huggingface.co/QuantStack/Wan2.2-Animate-14B-GGUF → ComfyUI/models/diffusion_models/

Quant	VRAM
Q4_K_M	~10–12 GB
Q5_K_S	~14–16 GB
Q6_K	~20 GB+

lightx2v I2V LoRA: https://huggingface.co/Kijai/WanVideo_comfy/resolve/main/Lightx2v/lightx2v_I2V_14B_480p_cfg_step_distill_rank64_bf16.safetensors → ComfyUI/models/loras/

WanAnimate relight LoRA: https://huggingface.co/Kijai/WanVideo_comfy/resolve/main/LoRAs/Wan22_relight/WanAnimate_relight_lora_fp16.safetensors → ComfyUI/models/loras/

Text Encoder: https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/text_encoders/umt5_xxl_fp8_e4m3fn_scaled.safetensors → ComfyUI/models/text_encoders/

CLIP Vision: https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/clip_vision/clip_vision_h.safetensors → ComfyUI/models/clip_visions/

VAE: https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/resolve/main/split_files/vae/wan_2.1_vae.safetensors → ComfyUI/models/vae/

---

Quick Start

1. Load workflow in ComfyUI
2. Upload reference image + input video
3. Adjust prompts
4. Set green (keep) / red (change) points in detection subgraph
5. Width and height must be multiples of 16
6. Run — saves automatically

---

Changelog

Version	Notes
v1.2	Looping, points editor out of subgraph
v1.0	Initial release