LTX 0.9.8 MAX QUALITY: Latent Upscaling & Tiled Sampling PRO Workflow (ComfyUI Tutorial)

This ComfyUI workflow is for LTXV upscaling with latent upscaling and tiled sampling; it accepts image, text, and random-noise inputs and includes VAE decoding and video combination, but its final output format is not specified.

The graph includes LTXV conditioning, latent upscaling, tiled sampling, VAE decoding, frame interpolation, film grain, text overlay, and video combination.

Named inputs include LoadImage, CLIPTextEncode, and RandomNoise.

Output-related nodes include VAEDecode, VAEDecodeTiled, and VHS_VideoCombine; the final output format is not specified.

The workflow is listed with the LTXV base model and publisher Lordcaocao2025.

The named workflow archive is ltx098MAXQUALITYLatent_v10.zip.

Listed node packs include comfy-image-saver, comfyui-frame-interpolation, comfyui-kjnodes, and ComfyUI-LTXVideo.

Other listed node packs are comfyui-videohelpersuite, comfyui-gguf, and comfyui-textoverlay.

Listed model files are LTX-Video-spatial-upscaler-0.9.8.safetensors, LTXV-13B-0.9.8-distilled-Q6_K.gguf, LTXV-13B-0.9.8-distilled-VAE.safetensors, and t5xxl_fp16.safetensors.

The model configuration is labeled 13B LTX-Video with GGUF, FP16, Q6, and Q6_K labels.

The workflow title identifies it as LTX 0.9.8.

Before setup, verify that all four named model files are available and confirm which packages provide comfyui-gguf and comfyui-textoverlay.

Suggestion · not verified

Confirm the output format you need because the workflow does not specify one.

Suggestion · not verified