Generation Workflows
Text to Video
This is the fastest workflow for exploration.
Recommended prompt structure:
- Subject
- Action
- Scene
- Lighting or mood
- Camera behavior
Example:
A premium sneaker on a reflective platform, slow rotating reveal, dark studio, crisp rim lighting, cinematic dolly-inImage to Video
This workflow depends heavily on the quality of the uploaded image.
Best practices:
- Use a clean subject with strong separation from the background
- Avoid cluttered compositions
- Start with subtle motion before trying aggressive camera movement
Audio to Video
This workflow currently requires LTX 2.3 Pro.
Best practices:
- Upload clear audio with a stable beat or obvious dynamics
- Keep the visual concept consistent with the audio energy
- Start in
1080pbefore scaling up other settings
Parameter Notes
Duration
- Longer duration increases credits
- Fast can support longer runs
- Pro is better kept for validated ideas
Resolution
1080pis best for first-pass testing2Kis a quality upgrade after concept approval4Kshould be reserved for final outputs
Camera Motion
Use motion intentionally:
dolly-infor product revealspan-leftorpan-rightfor scene scanningtilt-upfor vertical emphasisnonewhen the subject motion should stay dominant