Stable Audio 3 Small Sfx
Text to audio
From $0.04 / run
Text to audio models
A text to audio model makes sound from a description: a jingle, a background track, a song with lyrics, the sound of rain on a window or a spoken dialogue. You describe the genre, mood, instruments or the sound itself, and the model makes it. 25 models do this job in the catalog right now. Each model's page shows its form and a price guide in US dollars.
Models
Questions
- How is text to audio priced?
- Most of these models charge per track or by the length of the audio. Each model page shows its price guide in US dollars, and the studio shows the exact price for your settings before you generate. If a generation fails, the reserved amount goes back to your balance.
- What should the description say?
- For music: the genre, mood, tempo, instruments and, if the model sings, the lyrics. For a sound effect: the source, the space it is in and how long it lasts. The model page lists the length and other fields it accepts.
- Which model should I start with?
- Music, sound effects and speech are usually different models, so choose by the job first. Then play the examples on this page and try your description on the two closest.