ComfyUI-Wan-SVI2Pro-FLF Introduction
ComfyUI-Wan-SVI2Pro-FLF is an innovative extension designed to enhance the video generation capabilities of ComfyUI by integrating SVI 2 Pro motion continuity with Wan 2.2's First/Last Frame (FLF) style control. This extension allows you to create smooth and seamless videos from a sequence of frames, offering precise control over the start and end of each video clip. It is particularly useful for AI artists looking to produce high-quality, continuous video content without the technical hassle of manually stitching frames together.
How ComfyUI-Wan-SVI2Pro-FLF Works
At its core, ComfyUI-Wan-SVI2Pro-FLF works by combining two powerful techniques: SVI 2 Pro's motion continuity and Wan 2.2's FLF control. Imagine you are creating a video from a series of images. SVI 2 Pro ensures that the motion between these images is smooth and continuous, much like a director ensuring that each scene flows naturally into the next. Meanwhile, the FLF control allows you to lock the first and last frames of each segment, giving you the ability to dictate exactly how each clip starts and ends. This combination results in videos that are not only fluid but also precisely controlled at key points.
ComfyUI-Wan-SVI2Pro-FLF Features
- Motion Continuity: Ensures smooth transitions between frames, maintaining the flow of motion throughout the video.
- First/Last Frame Control: Allows you to lock the beginning and end of each video segment, providing precise control over the video narrative.
- Flexible Workflow: The extension supports workflows for up to 7 clips, but you can easily extend this by duplicating generation groups.
- Optional End Samples: You can choose to use or bypass the end_samples input, giving you flexibility in how you control the end of each segment.
- Adjustable Overlap: If you notice any ghosting or doubling at the seams, an overlap slider is available to help you achieve a clean stitch.
ComfyUI-Wan-SVI2Pro-FLF Models
The extension is compatible with models that handle standard SVI 2 Pro. However, for optimal results, it is recommended to use the finetuned models specifically designed for this setup. These models include:
- SFW Models:
- High Noise Model
- Low Noise Model
- NSFW Models:
- High Noise Model
- Low Noise Model
What's New with ComfyUI-Wan-SVI2Pro-FLF
- February 20, 2026: Introduced optional end_samples input, allowing for more flexible video generation.
- February 20, 2026 (Update №2): Added a workflow to continue video generation from an existing video, using either a prompt or the next last frame.
- February 21, 2026: Fixed a bug that sometimes caused the last frame to be ignored during generation.
- March 02, 2026: Added a switch to enable/disable the end_samples feature directly, enhancing user control over the First/Last Frame functionality.
Troubleshooting ComfyUI-Wan-SVI2Pro-FLF
- Ghosting or Doubling at Seams: Adjust the overlap slider in the Image Batch Extend With Overlap node to achieve a clean transition.
- Washed Out Frames: Increase the contrast slightly using the Adjust Contrast node before the Image Batch Extend With Overlap node.
- Prompt Not Followed: Increase the clip length to allow the model more time to adapt to new directions, or rephrase the prompt to encourage earlier action.
Learn More about ComfyUI-Wan-SVI2Pro-FLF
For further learning and community support, consider exploring the following resources:
- SVI Tutorial (Video)
- Stable Video Infinity Main Page
- KJNodes for ComfyUI These resources provide valuable insights and tutorials that can help you make the most of the ComfyUI-Wan-SVI2Pro-FLF extension.