Experience the power of LTX-2 (LTX Video). Native 4K resolution, synchronized audio, and optimized for consumer GPUs. The future of AI video is here and it's open source.
Create stunning videos from text or images using the LTX-2 model.
Your masterpiece appears here
Select your settings on the left and hit generate to see the magic happen.
Explore the capabilities of the LTX-2 model with these stunning features.

Generate high-quality videos from simple text prompts with LTX-2's advanced understanding.

Bring your static images to life with fluid motion and synchronized audio.

Bring your images to life by transferring motion from any video.

Transform your videos into breathtaking native 4K resolution with LTX-2.

LTX-2, also known as LTX Video, is the latest breakthrough in open-source AI video generation from Lightricks. Released in January 2026, it sets a new benchmark for performance and quality. Unlike closed models, LTX-2 provides full access to model weights and training code, empowering developers and creators to build, fine-tune, and innovate without restrictions. It generates native 4K video with synchronized audio, all while being efficient enough to run on consumer hardware.
LTX-2 offers a unique combination of high-end performance and accessibility.

Getting started with LTX-2 is simple, whether you are a developer or a creator.
Start with a detailed text prompt describing the scene, motion, and atmosphere. Or, upload an initial image to guide the video generation.
Adjust parameters like resolution (up to 4K), frame rate, and guidance scale to fine-tune the output to your creative vision.
Run the model to generate your video. You can further refine the results using keyframe control or by iterating on your prompt.
Deep dive into the technical capabilities that make LTX-2 a game-changer.
Generate videos directly in 4K resolution without the need for upscaling artifacts. Enjoy crisp, detailed visuals suitable for professional workflows.
LTX-2 excels at understanding complex prompts, effectively translating nuances in language into accurate visual representations.
Designed with efficiency in mind, LTX-2 can run on standard consumer GPUs, democratizing access to high-end video generation tools.
Seamlessly integrate LTX-2 into your existing workflows with full support for ComfyUI, enabling complex node-based generation pipelines.
Adapt LTX-2 to your specific style or subject matter. The model supports LoRA and full fine-tuning, giving you the flexibility to create custom models.
Get faster feedback on your generations with optimized inference speeds, allowing for a more iterative and creative process.
Start generating stunning 4K videos today with the power of open source AI.