A stylish, modern scene depicts a woman in black pin-striped attire and purple hair, conversing on her cell phone in front of an iconic red telephone booth. The background is slightly blurred to emphasize the subject's vibrant appearance. She wears sunglasses that add a chic touch to her look, complemented by her sleek black dress and polished accessories. The image conveys a sense of modern sophistication with its clean lines, vivid colors, and intriguing composition.
AI video workflows and models are everywhere and it might be hard to select the one for you. If you're reading this you're already on your path to using one of the top open source models currently.
LTX can deliver really powerful results in amazing speed and that's just the power of LTX - speed. With this video2video workflow we'll be transforming input videos in AI counterparts with amazing efficiency. Whether it's for quick visualizations or as an input for further generations, LTX is a great tool to have in your toolbox.
Verified to work on ThinkDiffusion Build: Feb 27, 2025
ComfyUI v0.3.18 with LTX v0.9.1 model support.
Why do we specify the build date?
ComfyUI and custom node versions that are part of this workflow that are updated after this date may change the behavior or outputs of the workflow.
We've also written another guide on Image to Video with LTX:
Phu Ngo
Comparison of LTX from Hunyuan, SkyReels and Wan2.1
How to run LTX in ComfyUI for Video2Video Generation
Installation guide
Custom Nodes
If there are red nodes in the workflow, it means that the workflow lacks the certain required nodes. Install the custom nodes in order for the workflow to work.
- Go to ComfyUI Manager > Click Install Missing Custom Nodes
- Check the list below if there's a list of custom nodes that needs to be installed and click the install.
Models
For this guide you'll need 4 models. 1 is pre-loaded and 3 will have to be downloaded manually.
1. ltx-video-2b-v0.9.1.safetensors
2. t5xxl_fp16.safetensors
3. Mistral-7B-Instruct-v0.3.Q4_K_M.gguf
4. florence2_model (base) - Pre-Loaded Model
- Go to ComfyUI Manager > Click Model Manager
- Search for the models above and when you find the exact model that you're looking for, click install and make sure to press refresh when you are finished.
Optional Model Path Source
Use the model path source if you prefer to install the models using model's link address and paste into ThinkDiffusion MyFiles using upload URL.
| Model Name | Model Link Address | ThinkDiffusion Upload Directory |
|---|---|---|
ltx-video-2b-v0.9.1.safetensors |
.../comfyui/models/checkpoints/ |
|
| t5xxl_fp16.safetensors | .../comfyui/models/clip/ |
|
| Mistral-7B-Instruct-v0.3.Q4_K_M.gguf | .../comfyui/models/llm_gguf/ |
|
| florence2_model (base) |
Pre-Loaded
|
Pre-Loaded |
Step-by-step Workflow Guide
This workflow was pretty easy to set up and runs well from the default settings. Here are a few steps where you might want to take extra note.
| Steps | Recommended Nodes |
|---|---|
| 1. Load a Video Load a video with a subject. Subjects closer to the center works best. |
 |
| 2. Set the Models In LLM node, Be sure to use a GGUF model. Otherwise, it wont appear in the Searge LLM node. |
 |
| 3. Write a Prompt Write a prompt at the top most node in this step. Describe what change you want to see in your video. The 2 other prompts are auto-generated from the LLM nodes. |
 |
| 4. Check the Generated Video Check the result of generation. If there's a problem with your result, just run the generation again. |
 |
| Set the Size (Optional) This can be found in Input Video Settings. If you're fine with the original aspect ratio you can use the keep_proportion setting. |
 |
| Change the Sampling Settings (Optional) Change these settings to finetune your results in the LTX Flow Edit Sampler. Skip steps range 5-15 Refine steps 20-25 LTXVScheduler steps 28-32. |
 |
Examples
A captivating video portrays a woman adorned in a white jacket and hat, standing amidst the serene winter landscape. Surrounding her are vehicles on the road, trees, buildings with windows, a light pole, and a railing. The snow-covered ground adds to the peaceful atmosphere as she stands gracefully against the backdrop of the visible sky. A subtle yet striking change has transformed the original male character into a beautiful woman, showcasing femininity while maintaining the essence of the original scene.
A whimsical scene featuring a fluffy, playful rabbit hopping gracefully across a lush green field. The surroundings are dotted with trees in the distance and vibrant wildflowers beneath its feet. The rabbit's long ears twitch as it bounds through the grassy expanse, its white fur standing out against the verdant backdrop. This digitally crafted illustration showcases an adorable twist on the original concept by transforming a white cat into a mischievous-looking rabbit, further enhancing the whimsical and charming atmosphere of the scene.
Depicts a military tank driving down a highway, with the ocean and towering trees on one side, buildings and a clear blue sky on the other. The dominant subject of the image is an armored military vehicle, now transformed into a sleek, powerful Honda CR-V-inspired tank design. The tank features reinforced armor plating, cannon turrets, and camouflage paint to blend in with its surroundings. Its treads dig deep grooves into the asphalt of the highway, leaving behind a trail of debris. The trees rustle in the wind, and buildings cast long shadows over the road, while the clear blue sky provides an ominous backdrop for this scene. A sense of motion is captured through skillful use of light sources that accentuate the tank's rugged exterior, giving it both a menacing yet majestic presence on the battlefield.
A unique scene of a metal robot sitting at a desk in front of a window, clad in a black t-shirt and wearing an intricate watch. On the desk lies a book and a pen, while to his right stands another metallic figure. Through the window, we can see a plant and billowing curtains blowing gently in the breeze. The metal robot is intricately detailed with shimmering surfaces that reflect the surroundings and give life-like qualities to its mechanical form. Its eyes glow with an intense light, suggesting sentience, while its articulated limbs are poised comfortably at the desk. The image exudes a sense of quiet contemplation as the robots work together in this futuristic scene.
Featuring a brown bird perched atop a tree branch, its majestic form standing out against the blurred background. The bird is depicted in intricate detail with feathers of a deep brown and golden hue, showcasing the sharp edges and powerful features that characterize this regal creature. In contrast to the vibrant colors of the hummingbird in the original prompt, this eagle's color scheme creates an air of power and elegance befitting its status as one of nature's most iconic birds of prey. The tree branch is rendered with delicate details that enhance the sense of realism and atmosphere, while the background remains slightly blurred to maintain focus on the striking bird perched atop it.
If you’re having issues with installation or slow hardware, you can try any of these workflows on a more powerful GPU in your browser with ThinkDiffusion. While this tutorial focuses on image to video generation for LTX, we've got another one coming up soon for video to video generation for LTX.
If you enjoy ComfyUI and you want to test out creating awesome animations, then feel free to check out the tutorials below. And have fun out there with your videos!
Phu Ngo
Phu Ngo
Phu Ngo
Member discussion