Selects a range of frames from a video latent. Features: - Supports positive and negative indexing - Preserves batch processing capa

Wraps transformer blocks to be able to skip attention layers.

Audio-video embeddings processor: dual connectors + concat.

Custom type that matches any ComfyUI type for dependency inputs.

r""" A basic Transformer block. Parameters: dim (`int`): The number of channels in the input and output. num_attention_heads

19B: per-segment norm -> aggregate_embed -> 3840

22B: per-token RMS norm -> rescale -> aggregate_embed(s)

ARRI LogC3 (EI 800) HDR compression. Maps linear [0, inf) -> [0, 1] via the camera log curve, then scales to [-1, 1] for VAE input.

Configuration for model components.

Configuration for sampling parameters.

Configuration for spatial tile processing.

Video-only embeddings processor: single connector.

Encodes text prompts using the LTX Video API, returning CONDITIONING for LTX-2 models. This node replaces the local CLIP encoding step by se

Extended IC-LoRA guide node with per-guide attention strength control. Same as LTXAddVideoICLoRAGuide, but adds attention_strength and attent

Extended keyframe guide node with per-guide attention strength control. Same preprocessing as LTXVAddGuideAdvanced (CRF, blur, interpolation, cro

Concatenates two video latents along the frames dimension. Features: - Validates dimension compatibility - Handles device placement

Dilates a video mask spatially and/or temporally using max-pooling and thresholds the result.

GPU-accelerated sparse track renderer. Renders circles at a high reference resolution and downscales with bilinear interpolation so circle si

Enhance prompts using Gemma 3 model. Supports T2V and I2V modes.

Decompress HDR from VAE output, tonemap for preview, optionally save EXR. Place after VAE Decode in an HDR IC-LoRA workflow. Recovers linear HDR

Applies image conditioning to the first frames of a video latent. Features: - Takes existing latent and applies image conditioning -

Composites images with a green (#66FF00) background where mask is active. If the mask has a single frame it is broadcast to match the video lengt

Blend two images seamlessly using Laplacian pyramid blending with a mask.

Preprocesses masks for use with the LTXVideo model's latent masking. This node handles temporal mask processing by: 1. Validating mask d

Provide speaker identity context for audio generation. Patchifies the audio latent and attaches it as reference tokens on both positive and n

Applies multiple masks to a video latent. Features: - Supports multiple input mask formats (2D, 3D, 4D) - Automatically handles fewe

Interactive spline editor for drawing sparse motion tracks. Provides a canvas widget where users can draw and edit spline control points on t

Load LTXV Audio VAE with dependency input for sequential loading in low VRAM scenarios.