ComfyUI-LTXVideo
Custom ComfyUI nodes for enhanced LTX-2 video generation with unified control and efficient latent processing.
About ComfyUI-LTXVideo
ComfyUI-LTXVideo is a collection of custom nodes that extend ComfyUI to support the LTX-2 image-to-video diffusion model. Since LTX-2 is already integrated into ComfyUI core, this repository focuses on unlocking advanced capabilities through additional workflows and node implementations.
The extension enables users to leverage LTX-2’s multi-condition control (e.g., depth or edges) via a single LoRA, and improves inference efficiency by operating on downsampled latents—reducing memory usage without sacrificing output quality.
It is intended for users already running ComfyUI and seeking deeper control and optimization for video generation tasks.
Key Features
- Unified LoRA support for multiple control conditions including depth and edges.
- Downsampled latent processing to reduce VRAM usage during inference.
- Faster video generation while preserving visual quality.
- Workflow-ready nodes designed specifically for LTX-2 within ComfyUI.
- Seamless integration with ComfyUI’s existing LTX-2 core implementation.
Pros & Cons
Pros
- • Enables advanced control over LTX-2 without requiring model retraining.
- • Reduces hardware requirements via optimized latent handling.
- • Open and free to use as part of the broader ComfyUI ecosystem.
Cons
- • Requires prior setup of ComfyUI and familiarity with its node-based interface.
- • No official documentation beyond README; users must refer to the main LTX-2 repo for model details.
- • Lack of explicit license information makes redistribution or commercial use uncertain.