A video's sound effects are generated from what it shows.

The Sonilo Sound Effects 1.0 model generates realistic sound effects synchronized to the rhythm of a video. What this audio engine changes for post-production.

Specializing in audio models designed for video, Sonilo is adding Sound Effects 1.0 to its offerings on fal. The model produces realistic sound effects from two types of inputs. In video mode, it analyzes what is happening on screen, including the movement and pacing of the scene, and then returns a single audio track synchronized with the footage rather than a vague background noise. In text mode, users simply describe the desired sound—such as heavy rain on a windowpane or distant city traffic—with the duration remaining adjustable. A video-to-video variant applies the sound effects directly to the provided shot.

Sound Effects 1.0 addresses a recurring blind spot in AI video production: an image may look finished while still requiring manual sound work. Sonilo claims training on a fully licensed catalog, presented as safe for commercial use.

On fal, the exclusive API partner for the launch period, video-to-SFX generation is priced at $0.009 per second of audio produced, compared to $0.0018 per second for text-to-SFX.