Lightricks LTX 2.3 Audio To Video

Generate cinematic short videos from audio, with optional first-frame image and prompt guidance.

Model Overview

LTX 2.3 Audio-to-Video turns an input audio clip into a short video sequence. You can optionally provide a reference image for first-frame grounding and a prompt to steer style and motion.

Best At

• Turning voice, music, or sound clips into visual motion.
• Producing short concept videos tied to an audio track.
• Fast iteration when combining audio + simple visual direction.

Limitations / Not Good At

• Intended for short clips (about 2-20 seconds).
• Results depend heavily on audio quality and prompt clarity.
• Highly specific scene continuity may need multiple reruns.

Ideal Use Cases

• Music and sound-driven teaser visuals.
• Podcast, voiceover, or spoken-word visual prototypes.
• Quick social clips from an existing sound asset.

Input & Output Format

• Input: required audio_url; optional image_url, prompt, and guidance_scale.
• Output: generated video URI returned on response.

Performance Notes

• Using a reference image improves first-frame control.
• Lower prompt ambiguity usually improves motion consistency.
• guidance_scale can be tuned, or left at -1 to let defaults apply.