Fuser Apps are here 🚀
Fuser Apps are here. Free generations for the next month 💫
Let's GobyLightricks
Seamlessly extend, prequel, and loop cinematic video clips with fluid physical dynamics
Extend or prequel any video sequence in three deliberate steps—set your anchor frame, tune temporal positioning, and direct the trajectory with natural language.
Engineered with a 1:192 spatiotemporal compression ratio and diffusion transformer architecture, LTX Extend excels at temporal outpainting, fluid physics, and smooth camera continuity.
Explore how creators extend cinematic sequences, sustain complex fluid dynamics, and outpaint temporal boundaries across horizontal and vertical aspect ratios.
From narrative filmmakers drafting prequels to commercial directors expanding product reveals, see how creative teams deploy LTX Extend across production workflows.
Use the start frame setting to control temporal direction. Setting start_frame_num to 0 places your source clip at the beginning of the timeline to append new footage after it (a sequel). Setting it to 120 places your clip at the end to generate preceding footage (a prequel), while intermediate values like 24 insert the clip in the middle to generate smooth transitions on both ends.
LTX Extend specializes in cinematic, 35mm film-like aesthetics with deep shadows, soft organic textures, and natural lighting. It excels at rendering complex physical dynamics such as billowing smoke, flowing water, rushing fire, and moving fabric, alongside continuous tracking shots and steady camera sweeps.
The model is not suited for rendering sharp typography, legible text, or intricate geometric vector shapes. It also struggles with hyper-detailed static imagery, complex multi-person crowded scenes, and multi-minute continuous shots without cuts, where spatial layouts may warp over extended temporal changes.
Write descriptive natural-language paragraphs detailing camera movement, lighting, subject action, and motion vectors rather than comma-separated keywords. You can keep prompt expansion enabled to automatically enrich simple concepts, or disable it when you need strict adherence to exact scripted instructions.
LTX Extend supports 480p and 720p output resolutions in both 16:9 widescreen and 9:16 vertical aspect ratios, running on an efficient 2-billion parameter Diffusion Transformer with inference steps configurable from 2 to 50.
Seamlessly extend, prequel, and loop cinematic video clips with fluid physical dynamics