How to Turn Your Poems Into Songs with AI: A Step-by-Step Method for Poets Who Don't Make Music
I was getting my teeth cleaned this week, and my dental hygienist started telling me a story. Her mother has written poems for decades. Notebooks of them. Drafts in drawers, finished pieces saved in folders, words that lived only on the page. Then her mother discovered a tool called Suno and started turning those poems into songs. The hygienist described what it was like to watch her mother finally hear her own words sung the way she had always imagined them in her head when she wrote them. Decades of poetry brought to life in audio. Her mother cried the first time it worked.
She told me this story without knowing that I'm the founder of an AI music education company. Without knowing that I've written a book on Suno that's currently the most comprehensive professional guide on the market. Without knowing that I help thousands of people figure out how to use these tools. She was just sharing something that had genuinely moved her, in the kind of conversation that happens at a dental chair while she scaled my teeth.
That moment was clarifying. These conversations are happening in the world right now. Not in tech blogs or industry conferences. In dental offices, kitchens, family gatherings, book clubs. People who have written for decades are discovering that the words they've held privately can become something they can play back, share, and listen to. The technology has caught up to what they always wanted to do.
If you're reading this, you might be that person. You've written poetry for years. Maybe decades. The words sit in notebooks, in journals, in files on your computer, in your memory. You've thought about what they would sound like as songs. You've maybe even tried to sing them out loud, alone, to hear how they go. And the barrier has always been the same: you're not a musician. You don't play an instrument. You don't read music. You don't have a band or a collaborator. The path from poetry to song felt closed.
It isn't closed anymore. This post is how you walk through it.
What's Actually Possible
Let me be honest about what AI music tools can do and what they can't.
Tools like Suno can take written words and produce a song. Real vocals. Real instrumentation. Real arrangement. Full audio that sounds like music, because it is. You paste your poem, describe what you want it to sound like, and the tool generates a song. The output isn't a finished masterpiece on the first try. It's a starting point that didn't exist three years ago.
What the AI doesn't know is what your poem is supposed to feel like. It doesn't know whether the speaker is your grandmother or your younger self. It doesn't know whether this should be slow and contemplative or fast and defiant. It doesn't know which words carry the most weight. Those decisions are yours. Your job is to communicate those decisions to the AI clearly enough that the output represents your poem instead of randomizing it.
That's the whole skill. Not the technology, the communication. The poets who turn their words into songs they actually love are the ones who learned how to direct the tool. The poets who try once, get a mediocre result, and give up are the ones who treated the tool like a vending machine: put in coins, hope something good comes out.
You don't have to gamble. You can direct. In Unlock Suno, I call this moving from operator mode to director mode: you stop hoping the machine understands you and start giving it decisions it can actually use.
Here's how to make those decisions.
The Four Decisions Before You Generate
Before you open Suno and paste your poem, decide four things. Each of these decisions shapes what the song will be. If you skip them, the AI picks defaults, and the defaults probably aren't your poem.
1. The Genre
Your poem will sound different as a folk ballad than as a country song than as an indie rock track than as a piano-driven singer-songwriter piece. Same words, completely different worlds.
Pick the world before you generate. If you're not sure, ask yourself: when this poem comes alive in your head, what do you hear? A solo piano? An acoustic guitar? A full band? A voice with sparse instrumentation? Drums or no drums?
You don't need to know music theory to make this decision. You just need to know what the poem sounds like to you. If you've ever listened to a song and thought "this feels like my poem," that artist or genre is your starting point. You can name it directly: "in the style of acoustic folk," "like a contemporary singer-songwriter," "like classic country."
2. The Voice
Male or female. Older or younger. Warm or sharp. Trained or raw. Smooth or weathered.
The voice carries the meaning of the words as much as the words themselves. A poem about losing your father sounds different sung by a 25-year-old soprano than by a 60-year-old baritone with a slight rasp. Neither is wrong. They're different songs.
Decide whose voice this is. If the poem is in your voice, picture yourself or someone close to your age and gender. If the poem is in someone else's voice (a character, a memory, an imagined speaker), picture them. The AI can do most voice types. You just need to tell it which one.
3. The Mood
Music shapes meaning more than poetry alone does. The same poem can be wistful, angry, hopeful, defiant, resigned, peaceful, or aching, depending on what you decide it is. The AI will pick a default mood if you don't specify one, and the default may not be your poem.
Be specific about the feeling. Not just "sad" but "quietly sad, like an early autumn morning." Not just "happy" but "warm, like a memory you're glad to have." The more specific you are, the more the song will actually be your poem.
4. The Structure
Most poems aren't structured the way songs are. Songs typically have verses, a chorus that repeats, sometimes a bridge that takes you somewhere different before returning. Your poem may not have a chorus at all. It might just flow from beginning to end.
Before you paste your poem into Suno, decide: does this poem have a line or a phrase that should repeat as a chorus? Are there sections that work as verses? Are you going to use the poem exactly as you wrote it, or are you going to restructure it slightly to fit a song format?
There's no wrong answer. Some poems are stronger if you leave them as-is and let the music handle the structure. Some poems become more powerful as songs when you find a chorus inside them and let it repeat. You're the one who knows the poem.
The Poem-to-Song Prep Sheet
Before you generate, write this down. It takes five minutes and it changes everything about what you get back.
Poem title:
Genre world: (acoustic folk, country, indie rock, piano singer-songwriter, etc.)
Vocal character: (gender, age, tone, any specific feel)
Emotional temperature: (the specific mood, in your words)
Core repeating line, if any: (the line or phrase that should be the chorus, if the poem has one)
Structure choice: (leave as poem, reshape into verse and chorus, or hybrid)
One thing the song must not become:
That last line is the most important one on the sheet. It introduces constraint thinking, which is how directors work. Examples: "must not become upbeat." "Must not sound theatrical." "Must not bury the words." "Must not feel like pop-country." Naming what the song must not become protects the poem from the AI's default tendencies.
Once you have your prep sheet filled out, you're ready to generate.
How to Actually Generate the Song
Open Suno at suno.com. You can use the free tier to start. You'll need to create an account.
Choose Custom mode, not the basic "describe a song" mode. The basic mode gives the AI too much creative freedom for poetry. Custom mode lets you control the inputs.
In the Style Prompt field, write a short description of the song using your four decisions. Genre, voice, mood, and any specific instrumentation. For example: "Acoustic folk ballad, female vocal in her 50s with warm tone, wistful and contemplative mood, fingerpicked guitar and light piano." Don't write paragraphs. The Style Prompt is for description, not storytelling.
In the Lyrics field, paste your poem. If you decided to add a chorus or restructure, do that here. Use simple markers if you want structure: [Verse], [Chorus], [Bridge]. These tell Suno where the sections are.
Important: only put singable words in the Lyrics field. Use [Verse], [Chorus], and [Bridge] as section labels, but do not write things like [Chorus: big emotional full band]. Suno may try to sing those words. Production directions belong in the Style Prompt, never in the Lyrics field. This is the single most common mistake new users make. Your poem ends up with strange words sung in the middle of it because the AI treated your production note as a lyric.
Generate. Suno will produce two versions. Listen to both. Note what's working and what isn't.
This first generation almost certainly won't be the final version. That's normal. The first generation is feedback, not the destination.
What to Listen For After the First Generation
After the first two generations, don't ask "do I like it?" That's not a useful question. The useful questions are diagnostic:
Does the vocal age match the speaker of the poem?
Does the mood match the poem's emotional truth?
Did Suno turn the poem into the wrong genre?
Did the chorus repeat the right idea, or did it repeat the wrong line?
Are any lines rushed or swallowed?
Is there one section worth keeping, even if the rest needs to change?
Each thing that's wrong becomes the next adjustment. You're not asking the AI to read your mind. You're directing it through iteration, the same way a producer directs a singer in a studio. Each take teaches you what to ask for next.
When You Get Stuck
The hardest part of this isn't the tool. It's knowing how to describe what you want when you don't have music vocabulary.
You might know that the first generation isn't right but not know why. You might want a "warmer" voice but not know how to ask for one. You might want the song to "build" but not know how to communicate that. This is where most people give up. The tool isn't the problem. The communication is the problem.
For this exact moment, I built a tool called Fader. Fader is an AI Studio Manager trained on the methodology I teach about Suno specifically. When you don't know how to describe what you want, Fader walks you through it. You can describe what you're hearing in plain language ("the voice feels too young," "the song feels flat," "the chorus doesn't land") and Fader translates that into the specific changes to make to your Style Prompt or Lyrics. It's like having a producer next to you who knows the methodology and helps you direct the tool.
Fader is available standalone, included in the Premium Starter Kit, and included in Red Lab Access. You can find it at jgbeatslab.com/meet-fader.
A Note for Poets Who Care About Ownership
Your strongest creative contribution is the poem itself. The words came from your life, your memory, your craft. The AI is producing the audio. You are producing the meaning.
Keep your drafts. Document your edits. Save your Suno prompts. Don't treat the AI output as the whole work; treat it as a recording made from your composition. The safer creative path here is human-authored, AI-assisted production: your words, your decisions, your selection, your arrangement, your finishing work. That's the position I teach across the JG BeatsLab catalog, and it's the position that holds up best for poets who want their work to remain unambiguously theirs.
Going Deeper
This article gives you the first layer: how to make the four decisions that turn a poem into a song direction. Unlock Suno goes deeper. It teaches the full Suno Stack (the ten-layer model that explains why Suno does what it does), the Failure Diagnosis Framework (how to figure out what's wrong with a generation and how to fix it), persona engineering (how to build repeatable artist identity across multiple songs), Style Box architecture, structure tag usage without confusing the model, and the full workflow from first generation to finished track.
If you only want the book, start with Unlock Suno: The Complete Guide. If you want the book plus real-time help applying the method, Fader is the companion tool. If you want the full entry toolkit, the Premium Starter Kit at $67 includes Unlock Suno, Fader, the supporting books on releasing and promoting your music, five Genre Blueprints, the Red Lab Quick Start Kit reference cards, and the 3-Song Sprint Course.
You can find all of it at jgbeatslab.com.
The Bigger Picture
If you've written poetry for years and never had a way to turn it into music, you do now. The technology is real. The methodology is learnable. The output on day one isn't a finished record, but it's a real song made from your real words, and that's a starting point that didn't exist for any generation of poets before this one.
The journey from poem to song is shorter than the journey from no-music-experience to learning an instrument and recording an album. It's shorter than finding a collaborator and convincing them to set your words to music. It's shorter than learning to sing well enough to record your own voice. For the first time, poets don't have to choose between their words and their music. Both are possible.
The dental hygienist's mother figured this out. So can you.
—Josh Founder, JG BeatsLab
