LTX Workflow

LTX Workflow is an AI video generation tool that lets creators and developers produce text-to-video and image-to-video content using the LTX 2.3 model inside ComfyUI or directly in the browser for those without a local GPU.

Built on Lightricks' LTX 2.3 architecture, the platform supports T2V, I2V, LoRA fine-tuning, V2V composition, and motion-track workflows. A built-in VRAM Adapter lets users select their GPU memory (16 GB, 24 GB, or 32 GB+) and instantly receive a matched model recommendation — for example, the FP8-quantized distilled checkpoint for 16 GB cards such as an RTX 4080 — along with the correct ComfyUI workflow JSON to download. Optional spatial (x1.5 or x2) and temporal (x2) upscalers extend output resolution and frame count without re-running the full model.

For users who prefer not to manage local hardware, LTX Workflow offers a browser-based online generation mode. The ComfyUI Workflow Configurator walks through GPU VRAM, generation type, quality preference, video length (up to roughly 8 seconds), orientation, resolution tier, and frame rate, then outputs a ready-to-use workflow JSON. This makes LTX Workflow practical both for hobbyists exploring AI video generation and for technical teams integrating Lightricks models into custom ComfyUI pipelines.

LTX Workflow is primarily used for generating AI videos from text prompts or reference images using the LTX 2.3 model inside ComfyUI or a browser. It is also used by technical teams to configure and extend custom ComfyUI pipelines with LoRA fine-tuning, upscaling, and motion-track workflows.