Sonilo V1.1 Video to Video

Sonilo V1.1 Video to Video generates an original, perfectly synced, commercially licensed music soundtrack for any video. It analyzes the footage (pacing, mood, visual rhythm, cuts) and composes music that follows the action, then muxes it back into the clip.

What it does best

• Scores to the picture. Beats, swells, and drops land on cuts and key moments (for example, a beat drop the instant a car launches, or a hit on every freeze in a dance). Feed it clear visual beats and it will hit them.
• Keeps the voice, adds the music. With Keep Speech on, it isolates the original speech or vocals and blends the generated music underneath with automatic ducking. Ideal for talking-head UGC, vlogs, interviews, and crowd scenes: cheering, laughter, singing, and chanting are all preserved too.

What it keeps vs replaces

• It preserves human voice and vocals only. Non-voice sound effects (engine roar, fireworks, footsteps, hooves, pure nature ambience) are not retained: the music replaces them. If you need those sounds, keep them in your source and mix outside Sonilo.
• With Keep Speech off (default), the whole original soundtrack is replaced by the new music.

How to use it well

• Prompt is optional. Leave it empty to let Sonilo read the video and choose a fitting style, or set a style (genre, instruments, mood, energy) to steer it.
• Turn Keep Speech on whenever the video has dialogue, narration, singing, or a crowd you want to keep hearing.
• Samples (1 to 3) produces variations from the same source so you can pick the best fit.
• Start Offset and Duration let you score only a segment instead of the whole clip.
• Cost scales with the scored video duration.