Creator Spotlight: John Feliciano of whispertone — The Craft Starts Before the Model Does
Share
There is a version of the AI music conversation that starts with the software.
Which generator did you use? What prompt did you write? How many versions did it take? What model sounds best?
That is not where the story of whispertone starts.
It starts with a guitar.
More specifically, it starts with a musician who spent years learning what it feels like when your fingers finally make the sound you were trying to hear.
John F. Feliciano remembers learning the opening riff of Nirvana’s “Smells Like Teen Spirit” when he was young. There was something powerful about pressing a few strings, getting the rhythm right, and suddenly hearing a familiar sound come out of his own hands.
That feeling stayed with him.
Before AI music generators existed, before Suno, before whispertone, John was already doing what musicians have always done: learning, playing, writing, rehearsing, performing, getting things wrong and trying again.
Eventually, guitar stopped being something he was learning how to do and became part of who he was.
That distinction matters when you look at what he is building now.
Because whispertone is not the story of somebody discovering that software can make songs.
It is the story of a musician discovering a new way to get old ideas out of his head and into the world.
Start with the First Look: whispertone — When the Guitar Demo Doesn’t Have to Stay in the Room.
Before whispertone, there were bands
John started playing guitar young.
His first real performance was at a birthday party when he was around 15.
By his own account, he probably was not ready.
That did not matter.
They had fun, played anyway and walked away feeling like they were on top of the world.
Later came more serious bands.
With DE’FEKT in Puerto Rico, the performances got bigger and the rehearsals became more disciplined.
Then came Message to Venus.
That was when John says he really began feeling that he was good enough to take music seriously.
The band received encouraging feedback from fans and other musicians. They toured. They rehearsed. They put in the work.
And John experienced something that is difficult to reproduce alone: the chemistry of several musicians making something together.
You bring an idea. Someone else changes it. Another person hears something you missed. There is compromise. There is friction. Then sometimes it locks into place and the final song becomes better than what any one person originally imagined.
That kind of collaboration teaches things no software can teach you.
It also comes with a cost.
Schedules. Logistics. Egos. Money. Studio time. People being available at the same time.
Good songs waiting because life got in the way.
John learned both sides of making music in bands.
He learned what is special about being in a room with other musicians.
He also learned how many ideas never leave that room.
Some songs never make it out of your head
That became part of the reason whispertone exists.
John still had guitar riffs. He still had lyrics. He still had structures and musical ideas.
But not every idea naturally became a rehearsal, a studio booking or a finished band project.
That is one of the places where generative technology changed the equation for him.
Not because the technology gave him ideas.
Because it gave existing ideas somewhere to go.
His process starts before Suno.
John creates the guitar demo. The riff comes first. The structure is his. The lyrics are his. The direction is his.
Then the production process moves into Suno.
That difference is important.
There is a tendency in the AI music conversation to treat everything that touches a generative model as if the model suddenly becomes the creator.
That is far too simple.
The better question is: What existed before the AI entered the process, and what decisions did the human continue making afterward?
In John’s case, there is a lot happening on both sides of that line.
whispertone itself is a one-man project John founded in Puerto Rico in 2025 and releases independently through whispertone music. The catalog now stretches across rock, alternative, metal and lo-fi, with three albums, three EPs and more than 30 singles.
The AI does not get the final vote
One of the most revealing things John told me was how often he says no.
This matters more than people realize.
Generative tools can produce technically impressive results very quickly.
That does not mean the result belongs in your project.
If Suno takes the track too far away from the feel of his guitar demo, John rejects it.
If the guitar identity gets buried, he redirects it.
If the result stops sounding like whispertone, it does not matter that the generation might sound polished.
It is wrong for the project.
That is creative direction.
And it is one of the places where experience matters.
A person who has spent years playing music does not only hear whether something sounds “good.” They hear timing, dynamics, arrangement, feel, how a guitar should sit in the track and whether something has become too clean, too busy, too generic or too far removed from the idea that caused the song to exist in the first place.
John describes Suno as production infrastructure.
That language is useful.
The tool can help carry the song further.
It does not decide why the song exists.
AI did not replace the band experience
John is not pretending the new workflow gives him everything he had before.
There are things he misses.
The chemistry of musicians making something together in real time. The feeling of a band locking into a groove. The physical experience of being onstage with the bass hitting your chest. The crowd. The shared experience.
Those things are different.
He does not see whispertone as a digital recreation of his old bands.
It is another creative outlet.
For John, the benefit is simple: he can take an idea much farther before he needs someone else’s schedule, budget or studio.
That matters when the alternative is the song never being made.
whispertone is intentionally transparent about AI
John has also made a decision that more creators are going to have to make.
He does not hide the AI.
The project is AI-assisted. He says so.
The name whispertone stays lowercase by design, and the project has never tried to create a fictional story where the technology was not involved.
Some people do not like that.
Some traditional musicians have objections. Some listeners hear “AI” and stop listening before they hear the work.
John accepts that.
He would rather people reject the project knowing what it actually is than accept it because he polished the truth into something easier to sell.
Transparency will not make every critic happy.
That is not the purpose.
The purpose is giving listeners accurate information about the creative process so they can decide for themselves how they feel about the work.
Success means something different now
John’s definition of success has also changed.
The younger musician cared about many of the markers most musicians care about: playing shows, releasing records, getting recognition, being taken seriously and proving that you belonged there.
Today, he measures whispertone differently.
Streams are nice.
They are not the scoreboard.
He measures success by whether the project gives his creative side somewhere to go, and whether someone else gets the opportunity to connect with the music.
That does not mean numbers no longer matter.
Anyone releasing music publicly understands that audiences matter.
But there is a difference between wanting people to hear your work and allowing platform metrics to determine whether creating it was worthwhile.
That is a distinction a lot of creators could benefit from making.
Promotion has been another experiment
whispertone has also been learning what happens after the song is finished.
John has largely focused on free promotional routes: internet radio, curator submissions, editorial outreach and music discovery opportunities that do not require paying for placement.
He has seen positive results through radio and editorial consideration, including airplay through Radio Wigwam and OTAT247/Mitxoda replaying bliss, while also discovering something many independent artists eventually learn: the music promotion industry contains a lot of people who are happy to take your money.
Paid playlist pitches. Paid PR. Pay-to-play opportunities. Services promising exposure.
John has generally walked away from those.
That does not mean paid promotion is automatically bad.
There are legitimate marketing services.
But creators need to understand the difference between advertising, professional promotion and paying somebody simply because they promise access to an audience.
John’s experience so far is that free outlets willing to consider AI-assisted music do exist, but they are still relatively sparse. Honest disclosure helps him identify which opportunities actually fit, while paid pitch-backs are common enough that he has learned to walk away from them.
Being transparent about AI can close some doors.
It can also help you find the doors that actually fit.
reflection brings the workflow into focus
John’s newest releases make the whispertone process particularly easy to understand.
In September 2026, he released two instrumental lo-fi EPs: reflection, vol. 1 on September 18 and reflection, vol. 2 on September 25.
Together they collect ten tracks originally released as singles between November 2025 and May 2026. Volume one brings together introspect, clarity, perspective, optimist and delight. Volume two collects melancholy, waves, journey, drift and vibe. A third volume is planned.
For John, reflection is about taking a moment to grow by thinking about past experiences and taking in the different phases of life. The first volume moves from looking inward toward optimism and delight; the second sits more with feelings that arrive, pass and leave something behind.
Every one of those pieces began with a guitar demo John wrote and recorded himself.
Then he developed and finished them through his AI-assisted production process, drawing on electronic textures, instrumental guitar, rock and Caribbean musical elements.
That detail matters because instrumental music strips away one of the easiest distractions in the AI debate.
You can follow the creative line clearly.
A guitarist creates an idea. The idea becomes a demo. The demo becomes the foundation. The production technology helps turn that foundation into a finished release.
John has also created one-hour versions of the pieces for YouTube, positioning the music for studying, driving, concentration and background listening.
It shows something else about where independent creators can go now.
A release no longer has to exist in one format.
A track can become part of an EP, then a long-form listening experience, then a video, then another piece of the larger catalog. Two of the tracks, drift and vibe, also appear on whispertone’s 2026 album enough.
Where whispertone goes next
John plans to keep releasing across two lanes: lyric-driven rock and alternative music, and instrumental lo-fi and ambient work. He also uses AI video tools for shorts and animated music videos that expand the songs visually rather than treating the audio release as the end of the idea.
That same approach shapes how he is trying to grow the project: more free airplay and editorial coverage, continued transparency about the hybrid process, and a catalog that can move between songs, long-form listening and visual storytelling.
The most important thing John said to AI-first creators
I asked John what he would tell someone discovering music creation through AI before they had ever learned an instrument.
His answer was simple.
Pick one up.
Learn something that makes a sound because your hands made it happen.
Wood. Metal. Nylon. Strings. Keys. Anything.
The point is not that every AI creator must become a virtuoso.
The point is that learning an instrument teaches things a prompt cannot: timing, dynamics, feel, patience, arrangement, what happens when your fingers cannot yet execute what your brain hears, and the difference between technically correct and emotionally right.
“The craft still starts before the model does.”
That is where his story becomes bigger than whispertone.
One of the risks of generative technology is that it can make creation look easier than learning.
Those are not the same thing.
You can generate something impressive very quickly.
Learning why it works takes longer.
And learning why something almost works may be even more valuable.
John also believes traditional musicians can learn something from the new generation coming through AI: speed of experimentation.
You can test an idea quickly now. Try an arrangement. Change direction. Explore what happens when a concept moves into another style.
That does not replace musical craft.
Used properly, it can accelerate exploration.
The craft starts before the model does
That may be the simplest way to understand whispertone.
The AI is real. It is part of the process. John is not hiding it.
But the creative story did not begin when he opened Suno.
It began years earlier.
Learning guitar. Playing badly before playing well. Standing at a birthday party at 15 and discovering what performing felt like. Rehearsing repeatedly with bands. Touring. Learning how collaboration works when everybody agrees. Learning how difficult it becomes when they do not. Writing riffs that never became songs. Holding onto ideas because there was no practical way to finish them.
Then eventually finding technology that allowed some of those ideas to leave the room.
That history does not make every whispertone release automatically good.
Nothing does.
The listener still gets to decide that.
What the history does is tell us something important about authorship in the AI era.
The question cannot simply be: Was AI used?
We need better questions.
What did the creator bring into the process? What did they decide? What did they reject? What did they rewrite? What existed before the model? What remained under human control afterward?
And ultimately: Why did this person need to make this thing in the first place?
For John Feliciano, the answer seems pretty clear.
The ideas were already there.
Now they have somewhere to go.
Listen to whispertone
Website: whispertonemusic.com
Suno profile: whispertone on Suno
Spotify artist: whispertone on Spotify
YouTube: whispertone on YouTube
bliss video: Watch on YouTube
reflection, vol. 1: Spotify | Apple Music
reflection, vol. 2: Spotify | Apple Music
One-hour reflection playlist: Listen on YouTube
Creator Classification: AI-Assisted
John writes and records the original guitar material, develops the song structures and lyrics where applicable, and directs the musical identity of the project. Generative AI enters the workflow as part of production and development rather than replacing the originating creative work.
Where this fits in the Jack Righteous system
whispertone is a useful example of what happens when a creator does not begin with the tool.
John already had a sound. He already had musical experience. He already understood what he wanted the guitar to do.
The technology helped him move those ideas further.
For creators building their own AI-assisted music workflow, that is the bigger lesson.
Do not only learn how to generate.
Learn how to listen. Learn what you want. Learn why you reject one version and keep another. Build enough musical vocabulary that you can direct the technology instead of simply accepting whatever it gives you.
Because the strongest AI-assisted creators will not be the people who generate the most.
They will be the people who know what belongs.
Your next step: If you are trying to build that kind of creative direction into your own music, start with Find Your Sound. If you are already releasing AI-assisted work, use AI Rights for Creators to understand the rights questions that come with the workflow.
You can also browse more real creator stories in the Creator Spotlight series.
