LTX Video 2.3 is INSANE Now — Face Constancy + Audio + RTX Upscale (Free Workflow)

LTX Video 2.3 just got way more powerful. I built a full ComfyUI workflow with face constancy control, two-pass tiled upscaling, audio-visual generation, distilled LoRA acceleration, and optional RTX super-resolution — all in one drag-and-drop JSON.
This workflow generates up to 10 seconds of video (264 frames) at up to 1024×1344 resolution while keeping your reference face locked and consistent throughout.
🔗 Download the workflow → https://www.patreon.com/posts/new-workflow-ltx-159215261?utm_medium=clipboard_copy&utm_source=copyLink&utm_campaign=postshare_creator&utm_content=join_link

✅ What's in the Workflow

🧠 Face constancy via image-to-video inplace conditioning
🎬 Two-pass pipeline: first pass generation + tiled latent upscale
🎵 Full audio-visual support with per-pass audio strength control
⚡ Distilled LoRA chain (4 slots) for faster inference — thanks Lightricks!
🖼️ RTX Video Super Resolution post-processing (optional)
🎛️ Clean parameter group: resolution, frame count, conditioning strength, frame rate
🔢 Seed control for reproducible results
🛠️ Custom Nodes You'll Need

ComfyUI-LTXVideo
ComfyUI-KJNodes
rgthree-comfy
ComfyUI-VideoHelperSuite
ComfyUI-JWNodes
ComfyUI-MathExpression
ComfyUI-mxToolkit
ComfyUI-STGGuider
ComfyUI-RTXNodes (optional, RTX GPU only)
ComfyUI-TwoWaySwitch
#comfyui #LTXVideo #aivideo #aivideogeneration #stablediffusion #comfyuiworkflow #LTX23 #FaceConsistency #aiart #texttovideo #imagetovideo #generativeai

Видео LTX Video 2.3 is INSANE Now — Face Constancy + Audio + RTX Upscale (Free Workflow) канала Aitotts