AlexandClaude Opus 4.5
08c6702b36
Enhanced LTXVImgToVideoInplace with: - Arbitrary frame_idx (negative values for last frame) - Chainable noise masks via get_noise_mask() - Preprocessing: CRF, blur, interpolation, crop Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
ComfyUI-LTXVideo-Extra
Extra nodes for LTX-2 video generation in ComfyUI.
Nodes
LTXV Img To Video Inplace At Index
Enhanced version of the built-in LTXVImgToVideoInplace that supports:
- Arbitrary frame placement — place an image at any frame, not just frame 0. Use
frame_idx=-1for the last frame. - Chainable noise masks — chain multiple instances to condition multiple frames (e.g., first + last frame) without masks overwriting each other.
- Preprocessing — CRF compression, Gaussian blur, interpolation method, and crop mode (matching
LTXVAddGuideAdvanced).
Example: First + Last Frame Conditioning with IC-LoRA
EmptyLTXVLatentVideo (97 frames)
→ LTXV Img To Video Inplace At Index (frame_idx=0, first frame)
→ LTXV Img To Video Inplace At Index (frame_idx=-1, last frame)
→ Add Video IC-LoRA Guide (driving video)
→ LTXVCropGuides
→ CFGGuider + SamplerCustomAdvanced
Install
cd /path/to/ComfyUI/custom_nodes/
git clone https://github.com/domprosys/ComfyUI-LTXVideo-Extra.git
Then restart ComfyUI. No additional dependencies required.
Update
cd /path/to/ComfyUI/custom_nodes/ComfyUI-LTXVideo-Extra/
git pull