Image Effects

This is the feature the product is named after. You upload a photo, and the model reads what's in it before generating audio that fits the scene. A street at night, a crowded kitchen, and a close-up of a broken hinge each pull a different result. It works best with a clear subject and decent light, since the model is reading visual content rather than file names.

Typing a description is the faster route when you already know what you want. Short, concrete prompts work better than long ones: "heavy wooden door closing in a stone hall" gives the model something to grip, while "sad and atmospheric" leaves it guessing. The same box accepts both inputs, so switching between them costs nothing.

Results arrive in seconds and play in the browser, which keeps the loop short enough to iterate. Most people run three or four variations before settling, and that habit is worth the credits. You hear the clip before you download it, so a bad take costs a credit but not a file on your drive.

Output length is the main thing the plans change. Free caps clips at 5 seconds, Starter at 10 seconds, and Premium at 20 seconds. For a UI click or a punch, 5 seconds is plenty. For ambience under a scene, 20 seconds is still short, and you'll be looping or layering the file in your editor.

Starter and Premium keep a record of everything you've generated, and the free plan doesn't. That sounds minor until a project comes back from review and you need the exact clip you made three weeks ago. Without history, that cue is gone and you're paying credits to rebuild something close to it.

The site sells royalty free sound effects, which is the practical reason to use a generator rather than a sample library with per-track licences. Read the terms before a commercial campaign, since royalty free covers the audio rights the site controls and says nothing about how you use the sound in context. For a YouTube video or an indie game build, the paperwork is simpler than a stock licence.