SUNO now transforms text into a spoken-word narration with background music

This is something many users have wanted for quite a while. I remember seeing posts at Discord where detailed instructions were given on how to get the AI to speak and not sing. Usually the AI broke into song somewhere in the generated piece, despite users' attempts to generate talk over music.

Now, all that has changed.

I tried a simple poem I wrote back in the 1980s. I was surprised by the result. Sort of a Leonard Cohen style narration that never breaks into singing. Exactly what I wanted.

This poem was written during a tough and uneven time. I'd just lost a partner of several years. The breakup left me sorrowful but it also unlocked a creative side that never went away. Suddenly I was writing songs, poems and painting watercolors in between my school work.

One of the watercolors was based on the Egyptian goddess Isis, whom I associated with a woman I had become infatuated with. She was distant and hard to reach, leaving plenty of room for me to project my own mythic beliefs onto her.

The result was this poem, originally entitled "Isis." Given the modern connotations now attached to that name, I decided to remove it and stick with the Suno-generated title, "Prismatic Eyes."

 © 2026 Michael W. Clark

Note - The image at Suno is not the original watercolor. It's created by Suno and nowhere as subtle as my original painting.