Quick Start

Let’s get started generating video with LTX-2.5.

Choose a Workflow

There are two ways to run LTX-2.5 locally. We recommend ComfyUI for the fastest way to go from zero to a generated clip using ready-made LTX-2.5 templates. If you’re integrating LTX into code or an automated pipeline, use the Python pipelines instead.

If you already have a plan in mind, start with the row that matches your input and goal. The Open Source guides describe local workflows; check the API documentation for hosted API requests.

ComfyUI Quick Start

ComfyUI provides an intuitive node-based interface with ready-made LTX-2.5 templates. If you haven’t installed ComfyUI yet, download it from comfy.org/download.

1. Install the LTX nodes

Recommended: ComfyUI Manager

1. Launch ComfyUI and open ComfyUI Manager
2. Search for “LTXVideo”
3. Click Install on the ComfyUI-LTXVideo node pack
4. Restart ComfyUI

Alternative: Manual installation

git clone https://github.com/Lightricks/ComfyUI-LTXVideo.git
cd ComfyUI-LTXVideo
uv sync --frozen
source .venv/bin/activate

For portable/embedded installations:

.\python_embeded\python.exe -m pip install -r .\ComfyUI\custom_nodes\ComfyUI-LTXVideo\requirements.txt

2. Generate your first video

1. Open ComfyUI’s template browser and pick an LTX-2.5 template (start with Text-to-Video if you’re new):
   • Text-to-Video — generate from a text prompt.
   • Image-to-Video — animate from a starting image.
   • First-Frame / Last-Frame — generate the motion between a start and end frame.
2. When prompted, download the model files the template needs. The Workflow Overview panel (or ComfyUI Manager) fetches anything missing.
3. Enter your prompt or add your references image(s).
4. Click Queue Prompt to generate.

Generation time depends on your GPU and VRAM.

Pipeline Quick Start

For programmatic use, run LTX-2.5 through the native ltx-pipelines package.

1. Install

git clone https://github.com/Lightricks/LTX-2.git
cd LTX-2
uv sync
source .venv/bin/activate

2. Download the models

LTX-2.5 ships as a split pack — one file per component. Grab the files needed; click Agree and Access on the model page on first download:

hf auth login
hf download Lightricks/LTX-2.5 \
  diffusion_models/ltx-2.5-22b-distilled-transformer-bf16.safetensors \
  text_encoders/gemma4-12b-with-proj-ltx-2.5-bf16.safetensors \
  vae/ltx-2.5-video-vae-bf16.safetensors \
  vae/ltx-2.5-audio-vae-bf16.safetensors \
  model_patches/ltx-2.5-duration-head-bf16.safetensors \
  latent_upscale_models/ltx-2.5-latent-spatial-upscaler-x2-bf16-1.0.safetensors \
  --local-dir models/ltx-2.5

3. Generate

uv run python -m ltx_pipelines.distilled \
    --transformer-path       models/ltx-2.5/diffusion_models/ltx-2.5-22b-distilled-transformer-bf16.safetensors \
    --text-encoder-path      models/ltx-2.5/text_encoders/gemma4-12b-with-proj-ltx-2.5-bf16.safetensors \
    --video-vae-path         models/ltx-2.5/vae/ltx-2.5-video-vae-bf16.safetensors \
    --audio-vae-path         models/ltx-2.5/vae/ltx-2.5-audio-vae-bf16.safetensors \
    --duration-head-path     models/ltx-2.5/model_patches/ltx-2.5-duration-head-bf16.safetensors \
    --spatial-upsampler-path models/ltx-2.5/latent_upscale_models/ltx-2.5-latent-spatial-upscaler-x2-bf16-1.0.safetensors \
    --prompt "A golden retriever running through a sunlit meadow" \
    --seed 42 \
    --output-path output.mp4

Every pipeline runs the same way from the command line — add --help to any module to see its options.

What’s Next?

Now that you have LTX-2.5 running, explore these guides to get the most out of the model:

Getting Help