
VFX
Keep control of your elements, colour, dynamic range, and detail as footage moves through compositing, grading, and finishing. LTX's open models give you outputs you can keep shaping and workflows you can adapt to your production.
Downloads on HuggingFace
Runs locally on 12GB VRAM
Near real-time generation
Why LTX
Separate subjects with soft alpha mattes and generate backgrounds guided by your 3D blocking. Keep control over how elements fit into the shot.
Preserve colour and dynamic range when editing EXR footage with Native HDR, or expand existing SDR footage into HDR. Keep room for exposure adjustments, compositing, and the final grade.
Improve clarity with Refine & Restore and use the Native Resolution workflow. Choose the tools your footage needs for the next stage of post-production.
Capabilities
Work at full delivery resolution on one GPU. No downscaling, no proxy round-trip, and barely any increase in VRAM consumption.

Rebuild fine detail with Refine, or bring archival footage back to clean, sharp, natural colour with Restore. Both scale to high resolutions using the Native Resolution workflow. Choose the tools your footage needs for the next stage of post-production.

Take any 8-bit SDR video input and turn it into 16-bit EXR in ACEScg. Grade it, keep working with it, or encode it for HDR-enabled screens.

Take 16-bit EXR in and get 16-bit EXR out through an ACES workflow. Color and dynamic range survive the edit, so the result stays ready for compositing, grading, and finishing.

Generate a polished render from 3D blocking, a prompt, or a reference image. Camera, composition, and frame alignment hold, so the result fits the shot you planned.

Generate alpha mattes with soft edges, partial transparency, and fine detail, including hair and smoke. Use the matte to control how an element blends with the layers behind it.

Security & IP
Adopt AI with the security, flexibility, and deployment options VFX teams require.
Deploy on-premises or in a private cloud.
Open weights for transparency and customization.
Self-hosted deployments ensure your assets never train shared models.
Explore
See the features that turn the model into a production pipeline. Built for scale, control, and integration.
Resources
See the features that turn the model into a production pipeline.