Run this workflow on InstaSD
Get started in minutes! Run this ComfyUI workflow online - no setup required.
Description
LTX-Video Workflow: Fast, High-Quality Video Generation with the 13B Model
LTX-Video is a powerful DiT-based video generation model that delivers high-quality, realistic videos at impressive speeds. Using the 13B model v0.9.7, this workflow generates 30 FPS videos at 1216ร704 resolution quicklyโideal for content creators, animators, and anyone needing efficient video generation. While itโs not real-time for most versions, the workflow still produces high-quality results in a fraction of the time compared to traditional video generation methods.
๐น Key Features:
- Text-to-Image & Image-to-Video โ Generate videos from text prompts or transform images into video sequences.
- Keyframe-Based Animation โ Define keyframes to create smooth animations and transitions.
- Video Extension โ Seamlessly extend videos forward or backward to create longer content from short clips.
- Video-to-Video Transformations โ Modify and enhance existing videos with new prompts and effects.
- 13B Model v0.9.7 โ This latest version offers fast video generation with incredible quality, perfect for iterative creative work.
- FP8 Version for Real-Time Generation โ The FP8 quantized model allows real-time generation on H100 GPUs, offering even faster performance with lower VRAM usage.
Models
| File | Destination | Source |
|---|---|---|
| t5xxl_fp16.safetensors | /ComfyUI/models/clip | Download |
| ltxv-13b-0.9.7-dev.safetensors | /ComfyUI/models/checkpoints | Download |
| ltxv-spatial-upscaler-0.9.7.safetensors | /ComfyUI/models/upscale_models | Download |
| ChatGPT Image May 20, 2025, 03_50_25 PM.png | /ComfyUI/input/3d | Download |
| ChatGPT Image May 20, 2025, 03_50_25 PM.png | /ComfyUI/input | Download |
Nodes
VAEDecodeTiledLTXVConditioningCLIPTextEncodeImageResizeKJVAEDecodeLTXVFilmGrainNoteSTGGuiderAdvancedKSamplerSelectLTXVLatentUpsamplerLTXVAdainLatentRandomNoiseSTGAdvancedPresetsSet VAE Decoder NoiseSplitSigmasCLIPLoaderCheckpointLoaderSimpleLTXVLatentUpsamplerModelLoaderLTXVTiledSamplerBasicSchedulerVHS_VideoCombineLTXVBaseSamplerLoadImage