LTX
Capabilities

Generate 2D and 3D animation styles, from flat illustration to cel-shaded cartoon, anime and storybook art.

LTX generates an alpha matte during the initial rendering pass, preserving hair, smoke, glass, and motion blur.

LTX-2.5 follows complex creative instructions from shorter, simpler prompts powered by a new Gemma 4 12B text encoder.

Generatively upscale low-resolution video with LTX-2.5's Pixel Spatial Upscaler IC-LoRA

Generates connected, multi-shot scenes in a single pass, holding character, environment, lighting, voice, and visual style across every cut.

Remove or replace objects, people, and regions in video with LTX's mask-conditioned inpainting IC-LoRA.

LTX-2.5 generates audio and video together, in sync — speech, foley, and ambience that land on the action, now cleaner with fewer artifacts.

Keep characters, props, and locations consistent across generated videos. Condition generations on a single sheet.

Fine-tune LTX-2.5's pretrained checkpoint into a world model for robotics, autonomous vehicles, and digital twins.

Apply a consistent look to a clip while preserving structure, identity, and motion.

Replace prompt roulette with locked control. Guide motion with trajectory splines, lock a camera move, or hold the structure of a reference.

Remove compression artifacts and sharpen blurred footage while preserving identity and framing.

Add realistic color to black-and-white footage, guided by prompt.

Remove, replace, or extend regions of a video while preserving the rest of the scene.

Continue any video from its last frame with a new prompt. No seams, no cuts.

LTX-2.5 brings HDR into the production pipeline, working with scene-linear EXR from input through output.

Controllable generative video for VFX teams: camera LoRAs, IC-LoRA transforms, 4K pre-vis.

Regenerate any moment in a video without re-rendering the whole clip.

Train custom LoRAs on LTX-2.3 for consistent characters, styles, or motion.

Generate video from text, image, and audio inputs combined for precise creative control.

Animate a still image into smooth, coherent video with natural motion.

Transform existing footage: restyle, retake, or regenerate segments while preserving motion.

Create cinematic video from text prompts with full control over motion and style.

Generate video directly from audio, sound drives motion, pacing, and scene structure.