Audio and voice
What the audio models do
Music models write an instrumental track from a text prompt. Sound effect models make short clips such as rain, footsteps or a door slam from a description. One of them generates sound to match a video. Voice conversion changes the voice in a recording you upload.
How audio is priced
Depending on the model, audio is priced per track, per clip or per 1,000 characters of text. The unit is shown next to each rate, and the form shows the price again before you generate.
Which audio model to start with
For background music, start with AI Music and describe the genre, tempo and mood in one line. Listen to the result before you build a longer edit around it. For short effects, Sound Effects works from a written description. MMAudio is the one to pick when the sound has to follow a video you upload.
To change the voice in a recording you already have, use Voice Conversion (RVC). It works on the file you upload. The words and timing come from your recording, and only the voice changes.
