Run this workflow on InstaSD
Get started in minutes! Run this ComfyUI workflow online - no setup required.
Description
Transform static images into vivid video animations with WAN 2.1 I2V Start/End Frame Workflow, enabling precise control over both starting and ending frames. This powerful method provides frame-anchored interpolation, ideal for storytelling, cinematic transitions, and animated scenes that maintain style, structure, and character integrity.
🎯 Features
- Start & End Frame Anchoring – Define the beginning and end visuals to guide the motion transformation.
- WAN 2.1 Integration – Leverages the strength of the WAN FUN model series for high-quality interpolation.
- Style-Coherent Transitions – Ensures visual and stylistic consistency between frames.
- Frame Count Control – Adjust the number of interpolated frames for pacing and duration flexibility.
- Built for ComfyUI – Entirely within ComfyUI with labeled, modular, and reusable nodes.
💡 Use Cases
- Image-to-Video Animation – Breathe life into static art with fluid motion between keyframes.
- Cinematic Scene Previews – Build animated cutscenes from concept art.
- Character Posing – Animate transitions between expressions or actions.
- AI Storytelling – Chain scenes using end-to-start frame logic to create coherent narrative videos.
⚙️ How It Works
-
Prepare Input Frames
- Select your start frame image and end frame image.
- Use
Get_img_startandGet_img_endnodes to feed them into the workflow.
-
Preprocess & Encode Style
- Optionally style both images using CLIP and
img_restyle. - Resize if needed to match WAN input specs (typically 512–768px square).
- Optionally style both images using CLIP and
-
Load WAN & Support Models
- Include
WAN FUN 1.3B,wan_2.1_vae.safetensors, and the appropriateCLIPandCLIP Visionloaders.
- Include
-
Interpolation via WAN
- Set the
num_frames(e.g., 20–60) to control transition length. - Use the WAN model's interpolation pipeline to create smooth in-between frames.
- Set the
-
Postprocess with RIFE (optional)
- Enhance fluidity with
RIFE VFIinterpolation if needed.
- Enhance fluidity with
-
Output Video
- Compile the generated frames into a video with
VHS_VideoCombine. - Adjust FPS, filename prefix, and compression to suit your use.
- Compile the generated frames into a video with
Credits: The-ArtOfficial
Models
| File | Destination | Source |
|---|---|---|
| Wan2_1-I2V-14B-480P_fp8_e4m3fn.safetensors | /ComfyUI/models/checkpoints | Download |
| wrapper_Wan2_1_VAE_bf16.safetensors | /ComfyUI/models/vae/wan | Download |
| wrapper_umt5-xxl-enc-bf16.safetensors | /ComfyUI/models/text_encoders | Download |
| wrapper_open-clip-xlm-roberta-large-vit-huge-14_fp16.safetensors | /ComfyUI/models/clip_vision/wan | Download |
| Wan2_1-I2V-14B-480P_fp8_e4m3fn.safetensors | /ComfyUI/models/diffusion_models | Download |
Nodes
WanVideoVAELoaderWanVideoTorchCompileSettingsLoadWanVideoClipTextEncoderINTConstantWanVideoDecodeImageResizeKJVHS_VideoCombineWanVideoClipVisionEncodeWanVideoSLGWanVideoImageToVideoEncodeWanVideoTextEncodeWanVideoBlockSwapWanVideoTeaCacheWanVideoSamplerLoadWanVideoT5TextEncoderLoadImageStringConstantMultilineWanVideoModelLoaderNote