KJNodes/ltxvgenerated

LTXV Add Guide Multi(LTXVAddGuideMulti)

Add multiple guide images at specified frame indices with strengths, uses DynamicCombo which requires ComfyUI 0.8.1 and frontend 1.33.4 or later.

LTXV Add Guide Multi

positive
negative
vae
latent
num_guides
positive
negative
latent
KJNodes

Description

The LTXV Add Guide Multi node adds multiple guide images at specified frame indices with per-guide strengths. It encodes each guide image using a video VAE and attaches the resulting conditioning keyframe information to both positive and negative conditioning pools. The guide latents are also appended to the input video latent, extending its length.

This node uses DynamicCombo, which requires ComfyUI 0.8.1 and frontend 1.33.4 or later.

Inputs

InputTypeDescription
positiveCONDITIONINGPositive conditioning to which the guide keyframe info will be added.
negativeCONDITIONINGNegative conditioning to which the guide keyframe info will be added.
vaeVAEThe video VAE used to encode the guide images.
latentLATENTThe video latent tensor. Guide latents are appended to the end of this latent.
num_guidesDYNAMICCOMBOSelects how many guide images to use. When this value changes, additional inputs appear dynamically for each guide: an image input, a frame index, and a strength value.

Outputs

OutputTypeDescription
CONDITIONINGCONDITIONINGPositive conditioning with guide keyframe data added.
CONDITIONINGCONDITIONINGNegative conditioning with guide keyframe data added.
LATENTLATENTThe input video latent with guide latents appended.

Usage Notes

  • Set num_guides to the desired number of reference images. The node will then expose a matching set of inputs:
    • Guide image (IMAGE) – the reference image to guide generation.
    • Frame index (INT) – the frame position in the video where this guide applies.
    • Strength (FLOAT) – how strongly the guide influences the conditioning at that frame (typically between 0.0 and 1.0).
  • The VAE should be the same one used to encode the video latents in your workflow.
  • Guide latents are appended to the end of the input latent, increasing the total frame count. Ensure downstream nodes account for the extended length.
  • This node is intended for the LTXV video generation model and lives in the KJNodes/ltxv category.

Requirements

  • ComfyUI version 0.8.1 or later
  • Frontend version 1.33.4 or later (required for DynamicCombo support)

Comments

Sign in with GitHub to join the discussion.

Loading comments…
LTXV Add Guide Multi (LTXVAddGuideMulti) - ComfyUI-KJNodes | ComfyUI Wiki