LTX 2.3 Add Audio to Existing Video by WW — Add Audio to Video

LTX 2.3 Add Audio to Existing Video by WW takes an existing video through LTX 2.3 video-to-video generation to add or replace audio, producing a reprocessed video whose visuals may be slightly altered.

It reprocesses the existing video rather than adding more video, with an option for LoRas.

The description says it seems to sync sound effects and ambient noises well.

The required node packs are ComfyUI-GGUF, rgthree-comfy, ComfyUI-KJNodes, and ComfyUI-VideoHelperSuite.

The WebP version additionally requires ComfyUI-WebpToPNGSequence.

Listed model files include amoral-gemma3-12B-v2-qat.Q8_0.gguf, LTX-2-19b-distilled-lora-384.safetensors, ltx-2-spatial-upscaler-x2-1.0.safetensors, LTX-2.3_audio_vae_bf16.safetensors, and LTX-2.3_video_vae_bf16.safetensors.

Additional listed files include LTX-2.3-distilled-Q8_0.gguf, LTX-2.3-spatial-upscaler-x2-1.1.safetensors, and taeltx2_3.safetensors.

You can add or replace audio, and the workflow includes an option for LoRas.

Use the notes to locate adjustable parameters; if the compact layout is hard to read, consider Comfy's link-hiding function.

Suggestion · not verified

This workflow assumes prior knowledge of how to run LTX 2.3; before use, confirm the listed model files are available and verify which ComfyUI-GGUF package it expects.

Suggestion · not verified

Treat lip-syncing as unsuitable unless the video has a very clear close-up of a character's face and the dialogue is already known.

Suggestion · not verified