What each setting does

Request

Describe a song, lyrics, stems, speech or a voice. The assistant can help you choose a workflow.

Sources

Choose music or voice recordings you own. A song can supply a continuation, replacement or separation step.

Exact text

Lyrics and speech text stay exactly as entered, including line breaks. Choose a new request when you want different words.

Choices

Listen to saved candidates and choose one before dependent steps continue.

Credits

The quote shows the most website credits you authorize. Actual spend may be lower; a reservation is not a final charge.

Title and style

The title names the saved work or voice. Style describes the musical or spoken direction; instructions add details to the request.

Music input and instrumental

Choose a description or images, or supply exact lyrics with a style. Instrumental requests music without sung vocals.

Audio region

For continuation choose a precise starting time; for replacement choose the start and end of the section. The quote requires an explicit region.

Separation mode

Choose the instrument or sound groups to extract. All 38 modes show their outputs; 12-track separation needs a generated song.

Voice and language

Choose a preset or your own registered voice and the speech language. The speech text is used exactly as entered.

Voice recording permission

Registering a voice reference requires your permission to use that recording. Registration is free; speech synthesis is a separately quoted paid step.