Image to Video with Wan Start & End Frame
Image to Video with Wan Start & End Frame
Used for image to video generation, defined by the first frame and end frame images.
Video/Image2Video_Wan_Start_End.json
Verified to work on ThinkDiffusion Build: April 1, 2025
Why do we specify the build date? ComfyUI and custom node versions that are part of this workflow that are updated after this date may change the behavior or outputs of the workflow.
Minimum machine size: ULTRA
Use the specified machine size or higher to ensure it meets the VRAM and performance requirements of this workflow.
Key Inputs
Load Image (Start Frame): Use any JPG or PNG showing your subject clearly to start your video
Load Image (End Frame): Use any JPG or PNG showing your subject clearly to act as the last part of your video. Make sure it's the same resolution as the load image.
Width & height: Optimal resolution settings are noted. Default is set to 480P, up to 72oP.
Prompt: as descriptive a prompt as possible.
Num Frames: Default set at 77, higher frame numbers may result in out of memory.
Frame Rate: 16fps is optimal for this Wan video model
Output Path
../outputs/Image2Video_Wan_Start_End/
Workflow
Examples
Notes to extend the clips:
If you want to extend your clip, right click on the last frame output above and click "Copy (Clipspace)" and go back to the start frame image loader, right click and "Paste (Clipspace)", from there you can adjust your prompt for what happens next in the clip, or include a new end image.
Doing this over and over without introducing an end image may result in deterioration in quality over time, and context may be lost.
If you want to combine your clips, bypass this current group and enable the multi video combine group below, load however many videos you have saved and queue again.