siliconflow/ComfyUI-LTXVideo

siliconflow/ComfyUI-LTXVideo

Description: LTX-Video Support for ComfyUI

Language: Python

License: Apache-2.0

Stars: 0

Forks: 0

Open issues: 0

Created: 2025-12-16T12:51:05Z

Pushed: 2025-12-17T00:44:16Z

Default branch: master

Fork: yes

Parent repository: Lightricks/ComfyUI-LTXVideo

Archived: no

README:

ComfyUI-LTXVideo

ComfyUI-LTXVideo is a collection of custom nodes for ComfyUI, designed to provide useful tools for working with the LTXV model. The model itself is supported in the core ComfyUI code. The main LTXVideo repository can be found here.

🚀 New to using LTXV with ComfyUI? See our Getting Started page

⭐ 16.07.2025 - LTXV 0.9.8 Release ⭐

🚀 What's New

1. LTXV 0.9.8 Model

The new model and its distilled variants offer improved prompt understanding and detail generation

👉 13B Distilled model

👉 13B Distilled model 8-bit

👉 2B from 13B Distilled model

👉 2B from 13B Distilled model 8-bit

👉 IC Lora Detailer

2. Autoregressive Generation Introducing new ComfyUI nodes that enable virtually infinite video generation. The new LTXV Looping Sampler node allows generation of videos with arbitrary length and consistent motion. ICLoRAs are supported as well—by providing guidance from existing videos (e.g., depth, pose, or Canny edges), you can generate long videos in a video-to-video manner. 👉 [Long Img2Video Generation Flow](#long-video-generation) 👉 [Long Video2Video Generation Flow](#long-video-generation)

3. Detailer ICLoRA Introducing the Detailer ICLoRA, which enhances generated latents with fine details by applying a few additional diffusion steps. This results in significantly more detailed generations. 👉 [Detailer ICLoRA Flow](#video-upscaling)

⭐ 8.07.2025 - LTXVideo ICLora Release ⭐

🚀 What's New in LTXVideo ICLoRA

1. Three New ICLoRA Models Introducing powerful in-context LoRA models that enable precise control over video generation:

• Depth Control: LTX-Video-ICLoRA-depth-13b-0.9.7
• Pose Control: LTX-Video-ICLoRA-pose-13b-0.9.7
• Edge/Canny Control: LTX-Video-ICLoRA-canny-13b-0.9.7

2. New Node: 🅛🅣🅧 LTXV In Context Sampler A dedicated node for seamlessly integrating ICLoRA models into your workflow, enabling fine-grained control over video generation using depth maps, pose estimation, or edge detection.

3. Example Workflow Check out [example workflow](#iclora) for a complete example demonstrating how to use the ICLoRA models effectively.

4. Custom ICLoRA Training We've released a trainer that allows you to create your own specialized ICLoRA models for custom control signals. Check out the trainer repository to get started.

⭐ 9.06.2025 – LTXVideo VAE Patcher, Mask manipulation and Q8 LoRA loader nodes. ⭐

1. LTXV Patcher VAE The new node improves VAE decoding performance by reducing runtime and cutting memory consumption by up to 50%. This allows generation of higher-resolution outputs on consumer-grade GPUs with limited VRAM, without needing to load the VAE partially or decode in tiles.

⚠️ On *Windows*, you may need to add the paths to the *MSVC compiler (cl.exe)* and *ninja.exe* to your system environment PATH variable.

2. LTXV Preprocess Masks

Preprocesses masks for use with the LTXVideo model's latent masking. It validates mask dimensions based on VAE downscaling, supports optional inversion, handles the first frame mask separately, combines temporal masks via max pooling, applies morphological operations to grow or shrink masks, and clamps values to ensure correct opacity. The result is a set of masks ready for latent-space masking. 3. LTXV Q8 Lora Model Loader

Applying LoRA to an FP8-quantized model requires special handling to preserve output quality. It's crucial to apply LoRA weights using the correct precision, as the current LoRA implementation in ComfyUI does so in a non-optimal way. This node addresses that limitation by ensuring LoRA weights are applied properly, resulting in significantly better quality. If you're working with an FP8 LTXV model, using this node guarantees that LoRA behaves as expected and delivers the intended effect.

⭐ 14.05.2025 – LTXVideo 13B 0.9.7 Distilled Release ⭐

🚀 What's New in LTXVideo 13B 0.9.7 Distilled

1. LTXV 13B Distilled 🥳 0.9.7

Delivers cinematic-quality videos at fraction of steps needed to run full model. Only 4 or 8 steps needed for single generation.

👉 Download here

2. LTXV 13B Distilled Quantized 0.9.7

Offers reduced memory requirements and even faster inference speeds. Ideal for consumer-grade GPUs (e.g., NVIDIA 4090, 5090).

*Important:* In order to get the best performance with the quantized version please install q8_kernels package and use dedicated flow below.

👉 Download here

🧩 Example ComfyUI flow available in the [Example Workflows](#example-workflows) section.

3. Updated LTV 13B Quantized version

From now on all our 8bit quantized models are running natively in ComfyUI, still with our Q8 patcher node you will get the best inference speed.

👉 Download here

⭐ 06.05.2025 – LTXVideo 13B 0.9.7 Release ⭐

🚀 What's New in LTXVideo 13B 0.9.7

1. LTXV 13B 0.9.7 Delivers…