ComfyUI-LTXVideo

AddLatentGuide

Compact description unavailable for this node.

DynamicConditioning

Clones a model and attaches a denoise-mask function that raises mask values to a step-dependent power. The adjustment can apply to only the first frame or to the full mask.

Lightricks/LTXV · 1 input · 2 parameters · 1 output

GemmaAPITextEncode

Encodes text prompts using the LTX Video API, returning CONDITIONING for LTX-2 models. This node replaces the local CLIP encoding step by sending the prompt to an external API for processing. It requires an API key and selects the text encoder from checkpoint metadata: ``model`` (LTX-2.5+) or ``model_id`` (LTX-2.3). Inputs: - api_key: Authentication key for the LTX Video API - prompt: Text prompt to encode - ckpt_name: Checkpoint file containing model metadata Returns: - CONDITIONING: Encoded prompt conditioning ready for LTX-2 video generation

Api Node/Text/Lightricks · 1 output

GuiderParameters

Compact description unavailable for this node.

ImageToCPU

Moves an input IMAGE tensor to the CPU and returns it.

Utility · 1 input · 0 parameters · 1 output

LowVRAMAudioVAELoader

Loads an LTXV Audio VAE checkpoint with dependency support.

LTXV/Loaders · 1 output

LowVRAMCheckpointLoader

Loads a diffusion model checkpoint with dependency support.

LTXV/Loaders

LowVRAMLatentUpscaleModelLoader

Load latent upsampler with dependency input for sequential loading in low VRAM scenarios.

LTXAddImageGuide

Compact description unavailable for this node.

LTXAddVideoICLoRAGuide

Adds one or more conditioning frames starting at a specified frame index and supports single images and multi-frame videos. It returns positive and negative conditioning together with a latent.

5 inputs · 9 parameters

LTXAddVideoICLoRAGuideAdvanced

Extended IC-LoRA guide node with per-guide attention strength control. Same as LTXAddVideoICLoRAGuide, but adds attention_strength and attention_mask inputs to control how strongly this guide's conditioning influences generation via self-attention.

6 inputs · 10 parameters

LTXAttentionBank

Compact description unavailable for this node.

LTXAttentioOverride

Compact description unavailable for this node.

LTXAttnOverride

Compact description unavailable for this node.

LTXFetaEnhance

Compact description unavailable for this node.

LTXFloatToInt

Compact description unavailable for this node.

LTXFlowEditCFGGuider

Compact description unavailable for this node.

LTXFlowEditSampler

Compact description unavailable for this node.

LTXForwardModelSamplingPred

Compact description unavailable for this node.

LTXICLoRALoaderModelOnly

Compact description unavailable for this node.

1 input · 2 parameters

LTXPerturbedAttention

Compact description unavailable for this node.

LTXPrepareAttnInjections

Compact description unavailable for this node.

LTXQ8Patch

Compact description unavailable for this node.

LTXReverseModelSamplingPred

Compact description unavailable for this node.

LTXRFForwardODESampler

Compact description unavailable for this node.

LTXRFReverseODESampler

Compact description unavailable for this node.

LTXVAdainLatent

Compact description unavailable for this node.

Lightricks/Latents · 2 inputs · 2 parameters · 1 output

LTXVAddGuideAdvanced

Adds a conditioning frame or a video at a specific frame index.

Conditioning/Video Models · 5 inputs · 6 parameters · 3 outputs

LTXVAddGuideAdvancedAttention

Adds a conditioning frame/video at a specific frame index with per-guide

Conditioning/Video Models · 6 inputs · 7 parameters · 3 outputs

LTXVApplySTG

Defines the blocks to apply the STG to.

Lightricks/LTXV · 1 input · 1 parameter · 1 output

LTXVAudioOnlyEmptyVideoLatent

Fixed minimal placeholder video latent for audio-only (text-to-audio)

Lightricks/Audio · 0 inputs · 0 parameters · 1 output

LTXVAudioOnlyModel

Run the LTX-2 audio/video model in audio-only mode for text-to-audio.

Lightricks/Audio · 1 input · 0 parameters · 1 output

LTXVBaseSampler

Compact description unavailable for this node.

Sampling · 7 inputs · 8 parameters · 3 outputs

LTXVDilateVideoMask

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

2 inputs · 2 parameters

LTXVDrawTracks

GPU-accelerated sparse track renderer. Renders circles at a high reference resolution and downscales with bilinear interpolation so circle sizes match the CPU version. All work — rasterisation, compositing and resize — stays on GPU.

0 inputs · 3 parameters

LTXVExtendSampler

Compact description unavailable for this node.

Sampling · 9 inputs · 5 parameters · 3 outputs

LTXVGemmaCLIPModelLoader

Compact description unavailable for this node.

Lightricks/LTXV · 0 inputs · 3 parameters · 1 output

LTXVGemmaEnhancePrompt

Enhance text prompts using Gemma 3 VLLM for improved video generation.

Lightricks/LTXV · 2 inputs · 5 parameters · 1 output

LTXVGetTilingSizes

Size calculator in the spirit of Comfy's Get Image Size, for tiled fusion.

Ltxvideo · 1 input · 6 parameters · 10 outputs

LTXVHDRDecodePostprocess

Decompresses VAE-decoded HDR (ACEScct or legacy LogC3), tonemaps for

Lightricks/HDR · 2 outputs

LTXVImgToVideoConditionOnly

Applies image conditioning to the first frames of an existing latent.

Conditioning/Video Models · 3 inputs · 2 parameters · 1 output

LTXVInContextSampler

Compact description unavailable for this node.

Sampling · 7 inputs · 1 parameter · 3 outputs

LTXVInpaintPreprocess

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 length. The fill is always SDR display-encoded ``#66FF00`` (``_BG_COLOR_RGB``), even when ``images`` are already ACEScct. Native HDR inpaint IC-LoRA was trained with that pixel-space key; converting the green into ACEScct changes the VAE latent cue and breaks inpainting.

2 inputs · 0 parameters

LTXVLaplacianPyramidBlend

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

3 inputs · 2 parameters

LTXVLinearOverlapLatentTransition

Compact description unavailable for this node.

LTXVLoadConditioning

Compact description unavailable for this node.

0 inputs · 2 parameters

LTXVLoadEXRSequence

Native HDR input: load EXR → ACEScct [0,1] frames for IC-LoRA / VAE encode.

Lightricks/HDR · 6 outputs

LTXVLoopingSampler

The LTXVLoopingSampler addresses two fundamental challenges in video generation: 1. Temporal Length: Generating very long videos by dividing them into overlapping temporal segments 2. Spatial Resolution: Creating high-resolution frames by splitting them into overlapping spatial regions Each "tile" is processed independently and then seamlessly blended with its neighbors to create the final coherent video output. - 🎬 Temporal Tiling: Generates long videos by processing overlapping time segments - 🖼️ Spatial Tiling: Creates high-resolution output by dividing frames into spatial regions - 🎯 Multiple Conditioning Modes: Supports image conditioning, guiding latents, and negative index conditioning - ⚡ Memory Efficient: Processes one spatial tile at a time to minimize memory usage - 🔄 Seamless Blending: Uses weighted blending for smooth transitions between tiles - 🎲 Advanced Seeding: Configurable per-tile seeding for reproducible results | Parameter | Type | Default | Range | Description | |-----------|------|---------|-------|-------------| | model | MODEL | - | - | The diffusion model for video generation | | vae | VAE | - | - | VAE for encoding/decoding between pixel and latent space | | noise | NOISE | - | - | Noise generator for the sampling process | | sampler | SAMPLER | - | - | Sampling algorithm (e.g., Euler, DPM++) | | sigmas | SIGMAS | - | - | Noise schedule defining the denoising steps | | guider | GUIDER | - | - | Conditioning guider (must be STGGuiderAdvanced to work with LTXVLoopingSampler) | | latents | LATENT | - | - | Input latent tensor defining video dim…

Sampling · 12 inputs · 12 parameters · 1 output

LTXVMultiPromptProvider

Compact description unavailable for this node.

LTXVNormalizingSampler

Compact description unavailable for this node.

LTXVPatcherVAE

Compact description unavailable for this node.

Lightricks/LTXV · 1 input · 0 parameters · 1 output

LTXVPerStepAdainPatcher

Compact description unavailable for this node.

Lightricks/Latents · 1 output

LTXVPerStepStatNormPatcher

Compact description unavailable for this node.

Lightricks/Latents · 1 output

LTXVPreprocessMasks

Preprocess masks to be used for masking latents in the LTXVideo model.

Lightricks/Mask Operations · 2 inputs · 7 parameters · 1 output

LTXVPromptEnhancer

Enhances text prompts for image generation using LLMs.

Lightricks/LTXV · 2 inputs · 2 parameters · 1 output

LTXVPromptEnhancerLoader

Downloads and initializes LLM and image captioning models from Hugging Face to enhance text prompts for image generation.

Lightricks/LTXV · 0 inputs · 2 parameters · 1 output

LTXVQ8LoraModelLoader

Compact description unavailable for this node.

Lightricks/LTXV · 1 input · 2 parameters · 1 output

LTXVSaveConditioning

Compact description unavailable for this node.

1 input · 2 parameters

LTXVSaveHLG

Writes a BT.2020/HLG 10-bit HEVC master from scene-linear hdr_linear

Lightricks/HDR · 2 inputs · 3 parameters · 0 outputs

LTXVSDRToHDRWorkingSpace

SDR → ACEScct working space for HDR IC-LoRA (pipelines SDRColorSpace).

Lightricks/HDR · 1 output

LTXVSelectLatents

Selects a range of frames from the video latent.

Latent/Video · 1 input · 2 parameters · 1 output

LTXVSetAudioRefTokens

Provide speaker identity context for audio generation. Patchifies the audio latent and attaches it as reference tokens on both positive and negative conditioning. The model prepends these tokens with negative temporal positions so they serve as identity context without being part of the generated output. Also outputs a frozen copy of the audio latent (noise_mask=0) for direct use in stage 2 without re-encoding.

3 inputs · 0 parameters

LTXVSetVideoLatentNoiseMasks

Applies multiple masks to a video latent.

Latent/Video · 2 inputs · 0 parameters · 1 output

LTXVSparseTrackEditor

Interactive spline editor for drawing sparse motion tracks. Provides a canvas widget where users can draw and edit spline control points on top of a reference image. Outputs interpolated track coordinates compatible with LTXVDrawTracks.

1 input · 3 parameters

LTXVStatNormLatent

Compact description unavailable for this node.

Lightricks/Latents · 1 input · 5 parameters · 1 output

LTXVTiledFusionSampler

Per-step tiled latent fusion for IC-LoRA (and other LTX conditioning) at resolutions larger than the model's trained window. This is not LTXVTiledSampler. That node independently samples a grid of tiles. This node keeps one latent canvas and one noise field, runs the model on overlapping crops at every denoise step, and Gaussian-blends the stepped tiles back onto the canvas. Tiles share a trajectory, so seams do not form. Peak VRAM tracks spatial tile size and the temporal window actually stepped (one content window plus one guide block when streaming), not the full canvas. Cached tile latents/noise/masks follow canvasdevice; processed conds stay on the GPU. Streaming also does one full-res VAE encode per temporal window. UNET + IC-LoRA Gemma CLIP → CLIPTextEncode (pos / neg) guide frames → LTX Add Video IC-LoRA Guide (positive, negative, latents) KSamplerSelect (euler / heun / …) → SAMPLER sigmas (Manual Sigmas or any scheduler, descending, ending at 0) │ ▼ LTXV Tiled Fusion Sampler → LATENT → VAE Decode (tiled) The latent must come from the IC-LoRA guide node. It carries appended guide frames and a noise mask. This node crops those extra frames itself, so LTXVCropGuides is not needed. Leave usestreaming off on the guide (the default) for a whole-clip encode. For clips longer than the trained window, turn usestreaming on and set the same tileframes (typically 97) on both the guide and this sampler. That re-encodes each temporal window as its own clip and writes ltxvtiledfusiontemporalplan onto the conditioning. The sampler disables windowing if the plan is missing. Do not…

Sampling · 7 inputs · 9 parameters · 1 output

LTXVTiledSampler

Compact description unavailable for this node.

Sampling · 8 inputs · 8 parameters · 2 outputs

LTXVTiledVAEDecode

Compact description unavailable for this node.

Latent · 2 inputs · 6 parameters · 1 output

LTXVVAEForceFloat32

Forces float32 on the video VAE (vae_dtype + first_stage_model) for

Lightricks/HDR · 1 input · 0 parameters · 1 output

ModifyLTXModel

Compact description unavailable for this node.

MultimodalGuider

Compact description unavailable for this node.

Set VAE Decoder Noise

Compact description unavailable for this node.

STGAdvancedPresets

Compact description unavailable for this node.

STGGuiderAdvanced

Compact description unavailable for this node.

STGGuiderNode

Implements Spatiotemporal Skip Guidance (STG), a training-free method enhancing transformer-based

Lightricks/LTXV · 3 inputs · 3 parameters · 1 output