Lyvia Lumi

About

Lyvia Lumi is a Pro-exclusive video model that generates video with synchronized audio. It works as text-to-video from a prompt alone, or image-to-video when you provide a start image. Supports durations from 4 to 10 seconds at 720p, with negative prompt support to exclude unwanted elements. One of the most affordable video models in the Lyvia lineup.

How to prompt

• Describe both the visual and the sound: "rain falling on a window, soft lo-fi music in the background"
• Use negative prompts to exclude artifacts: "blurry, distorted faces, text overlays"
• Provide a start image for image-to-video — the model animates forward from it
• Shorter durations (4–6s) tend to be more coherent; use 8–10s for slower, ambient scenes
• At 40 credits per 4s clip, Lumi is great for rapid iteration before committing to a higher-end model