How to Create Spoken Word in Musicfy With a Custom Voice: Two Real Experiments

JACK RIGHTEOUS CREATOR LAB · SEPTEMBER 30, 2026

Two real experiments: why Musicfy sang the first script—and how a rewrite produced spoken word

Can you use Musicfy to deliver a four-minute spoken-word monologue instead of singing it? I tested it twice using my Creator Era manifesto, I Wasn't Waiting for Permission. The first generation sang. The second, after I restructured the Custom Lyrics as natural prose and revised the Style instructions, produced a mostly spoken performance. After listening more carefully, I heard a few seconds of singing near the end—which I liked as an artistic touch—and a short phrase repeated twice. This is encouraging progress, not a finished or error-free result.

Independent creator testing: My observations describe these two outputs, not a guarantee that every Musicfy mode or account will reproduce them.

2. Why the first attempt became a song

The first Custom Lyrics included verse headings, a timed introduction, instructions about bass and percussion, and many one-line poetic fragments. Although my aim was spoken word, the text resembled song lyrics. That made the result a useful failed experiment.

I cannot isolate any single cause from two generations because I changed both Custom Lyrics and Style before the successful attempt. But I can show the workflow that worked in this test.

3. Rewrite Custom Lyrics as natural speech

Compare the two openings:

First version: lyric-style presentation

[INTRO — 0:00–0:25]
[Deep atmospheric bass. A distant piano.]
You know what's funny about dreams?
Everybody tells you to chase them.
But nobody talks about the price of admission.
The equipment you couldn't afford.

Revised version: narration

You know what's funny about dreams? Everybody tells you to chase them, but very few people talk about the price of admission.

Think about it. The equipment you couldn't afford. The people you couldn't reach. The opportunities that never came your way. And all those beautiful ideas that never made it out of your head.

Instead of adding more song labels, remove them. Use complete conversational paragraphs, straightforward punctuation and transitions such as “Think about it” or “Here's something I've come to understand.” Do not include musical instructions in Custom Lyrics when your intention is speech only.

4. The Style prompt that improved spoken-word delivery

Copy the following into the Musicfy Style field. Adapt the voice description to suit your own narrator.

SOLO SPOKEN-WORD AUDIOBOOK NARRATION.

Natural, unscripted conversational speech.

One mature male narrator with a deep, rich, gravelly Jamaican baritone and a natural Jamaican accent.

Speak exactly as an experienced storyteller would when recording a personal documentary or audiobook.

Use ordinary conversational sentence structures, natural breathing, authentic emotional expression and realistic changes in speaking pace.

Reflective, thoughtful and intimate at first. Gradually become more confident and inspirational.

Read the supplied text as continuous prose, not as poetry, lyrics or a musical performance.

Strictly spoken human speech from beginning to end.

No singing, chanting, melodic intonation, rapping, musical rhythm, harmonies or backing vocals.

No music. No instruments. No sound effects.

Clean, dry, isolated human voice recording.

Notice that this is a performance prompt, not a music-production prompt. It describes the speaker, speaking behavior and emotional direction. The Custom Lyrics separately provide the words as readable prose.

5. Reproduce the test

1. Write an original conversational monologue. Roughly 500–600 words is a starting point for about four minutes, depending on speaking speed and pauses.
2. Open Musicfy's song-creation workflow and locate Custom Lyrics and Style controls, where available.
3. Paste prose paragraphs into Custom Lyrics. Remove VERSE/CHORUS labels, instrumental directions and rhythmic line breaks.
4. Paste the narration prompt into Style. Change the voice attributes if needed.
5. Generate, listen, and preserve the result before changing anything.

Evaluate whether the voice actually speaks, whether it maintains its character, whether emotional emphasis makes sense and whether the entire script is delivered.

6. What this does—and doesn't—prove

In my test, replacing song-like Custom Lyrics with continuous prose and strengthening audiobook narration instructions changed the output from an unwanted sung interpretation to spoken word. The second output is mostly successful, not production-final: it includes a brief sung passage near the end that I enjoyed, plus an unintended short phrase repeated twice. Because I changed more than one variable, the comparison also cannot isolate which instruction caused the improvement.

A useful follow-up is to keep the revised script identical and vary only the Style instruction. A second follow-up can keep Style identical and alter only the script layout. That will tell us more about how much control Musicfy offers for narration.

7. Keep creating

Start with the Musicfy Creator Hub, then follow the Musicfy Create a Song Guide. For a different voice-production experiment, read the Musicfy Voice Cloning Tutorial.

Save your unsuccessful attempts. Sometimes the difference between the first and second version becomes the most useful lesson you can share.

Create What You Love | Love What You Create.
Gary Whittaker, Founder and Operator, JackRighteous.com