LTX-2 in Promptus

LTX-2 is a video generation model designed for controllable, high-quality clips. It supports common workflows like text-to-video and image-to-video, and is optimized for consistent motion, structure, and camera behavior across a shot.

Use Fast when you’re iterating and need quick feedback. Use Pro when you care more about stability, detail, and shot-to-shot consistency. A simple rule: Fast for exploration, Pro for finals.

Yes. LTX-2 is built for directed motion—meaning you can guide structure and movement using controls like pose signals (e.g., OpenPose) and other composition guides, then use prompts to define style, identity, and camera intent.

Yes. You can apply LoRAs to steer visual style, character identity, or motion behavior, and keep that “look” consistent across multiple clips. This is especially useful when you’re building a repeatable series, brand style, or character library.

Use a CosyFlow: it turns a full ComfyUI workflow into a clean, reusable interface with only the inputs you care about. That means you can rerun the same LTX-2 setup with new prompts, images, and settings—without rebuilding node graphs or breaking the workflow.