Add Spoken Word to AI Songs

Not every part of a song needs to be sung; spoken intros, narration, and spoken-word sections can add character and storytelling.

Spoken parts work best when they sound conversational, so read lines aloud and avoid stiff or overly complex wording.

Use labels such as Intro (spoken), Verse, Chorus, or Spoken word between sections so the model can understand the difference.

Use concise phrases that support the song rather than replacing the hook or overcrowding the arrangement.

Describe whether speech should feel intimate, cinematic, energetic, whispered, documentary-like, or calm.