← back to the directory
Model Inference & Serving

LTX-Video

1Fast 2diffusion 3transformer 4generates 5real-time 6video

the six Ws · specification

W1 Who

Lightricks, maker of the Facetune and Videoleap apps, released LTX-Video.

W2 What

LTX-Video is an open-weight video diffusion transformer optimized for speed, capable of generating video faster than it takes to watch on consumer GPUs.

W3 Where

Hosted on GitHub at Lightricks/LTX-Video and on Hugging Face under Lightricks.

W4 When

Released in November 2024.

W5 Why

It prioritizes real-time generation speed, making local, low-latency video generation practical for creators.

W6 With

Built on PyTorch and the diffusers library, with ComfyUI integration for local workflows.

W7 Watch

Released under a permissive custom Lightricks open license allowing broad commercial and non-commercial use of the weights; released directly by Lightricks.

text-to-videoreal-timediffusion transformerLightricks

for agents & scripts

Reading this as a machine? Query it directly.

Search is open JSON - no key. Report telemetry after using a tool and it feeds that tool’s Proof Score. Or speak MCP to /mcp and discover tools mid-loop.