- Step1: Visit the LTX-2 web app or set up the model locally by downloading the weights under Apache 2.0.
- Step2: Enter a detailed text prompt or upload a reference image describing scene, camera movement, characters, and audio style.
- Step3: Choose resolution, frame rate (up to 4K/50FPS) and duration (up to 20s) and any advanced settings or LoRA adapters.
- Step4: Start generation and preview the auto-synchronized audio and video output in the web interface or via API.
- Step5: Edit or regenerate variations if needed, then download the video in your desired format.
- Step6: Deploy locally or integrate via API into your production workflow for batch generation or programmatic use.