Nobody's Weatherman Built the Model
Your weatherman didn't build the model he forecasts from. He pulls the same public model runs anyone with a browser can pull — then bends them with what he knows about how they behave in his valley, in this season, on a setup that looks like this one. Nobody thinks less of him for it. That judgement is the job.
Ours works the same way. The engine that renders the audio is AI — we say so on every page, on every song, and we always will. What we've spent years building is the thing that steers it: the rules, the structure, the questions we ask you, and the way your answer becomes a lyric that couldn't have come from anyone else.
We'll show you the readings — why a slow dance sits at 74 BPM, why the pedal steel enters where the story softens — but the method stays ours.
Anyone can pull up the raw model run. Almost nobody can tell you what it's going to do.
Hear a forecast
Bring or Take
Reed Township
The brief was a marriage and a cooler argument. The forecast came back at 92 BPM with a banjo. Every choice traceable to the story.
Why you get more than one
Two versions of a song can share every word and still not be the same song. A snare that lands on two and four instead of holding back. A chorus that opens a bar early. A voice sitting closer to the microphone, so you hear the breath before the line. None of that changes what the song says. All of it changes whether it lands.
And the version that lands is not the one we would have bet on. It is usually decided by something small enough that nobody could call it in advance — least of all the person who built it.
So we do not guess on your behalf. We build the song more than once and let you hear the difference, because the gap between “that’s good” and “that’s the one” is almost never in the words. It is in the details, and the details are yours to choose. You keep every version either way.
The same song, more than one way
Within each group every word is identical. Nothing else is. This is the choice you get handed at the end — and it is the reason we do not make it for you.
One lyric. Three genres. Nothing else changed.
Alternative rock
Rock
Blues
Same lyric, same genre, three passes — and the third one comes back tender.
Pop · rowdy
Pop · rowdy
Pop · tender
Another song, two ways.
Grunge · hard rock
Rock
Anatomy of a Commission
We won't explain how the Song Architecture operates. But we can show you what it decided, and why — the evidence left behind in a finished song.
The story
"The brief: her Cavalier died on Highway 83, twenty below. He stopped. Three years later she married him."
A narrative ballad — verses advance the story chronologically, chorus lands the emotional turn.
A first dance is a slow conversation. 74 is where the body settles. G major for warmth without brightness.
Upright bass and kick move as one — a heartbeat the whole arrangement rests on.
The steel enters where the story softens, but it lifts instead of weeping — the memory is warm, not wistful.
One full bar of silence under the line "she married the man who stopped." The arrangement performs the title.
The song grew a second voice: every line Jenna actually says, she sings herself — first person, her own mouth. Pronouns match the mouth. That is a house rule.
The rules at work
First draft had a line the rules killed — 'gave her half my battery' — cute, but you can't give half a battery, and cute is how false sneaks past the ear. The rules rewrote it. That's the method working.
The record
I'd Stop Again
Five steps, one song
You Tell Your Story
Fill out the intake form with the details that matter — names, dates, memories, inside jokes, the feeling you want the song to capture. The more you give us, the more personal it gets.
We Listen
~MAK reads every word. Your story becomes a production direction — not a template, but a custom plan for a song that could only be yours.
The Song Architecture
Using a proprietary method we've spent years calibrating, ~MAK builds your song from the ground up — arrangement, instrumentation, vocals, lyrics. The method is ours. The song is yours.
You Choose
Every time we build your song you get six finished versions of it — six in one genre, or two each in three genres if you would rather compare directions. They are whole songs, not clips, and every one is yours to keep. What your tier buys is how many times we build it: once, twice, or three times, with you telling us what to change in between. Changes after you have settled on your final version are $50. Changing the song's language level — cleaner or stronger — counts as a standard revision.
Delivery
Your finished song arrives as a WAV master AND an MP3. The WAV is the studio file — the one thing we never sell off the shelf, at any price, to anyone who did not commission the song. The MP3 is the copy that travels: your phone, a text to your sister, the car. Both are yours, with your personal-use license.
Three lanes, one choice
Stock music libraries
- — Generic
- — Licensed
- — Already exists
- — No personalization
Good for background. Not for the moment.
Instant AI generators
- — Fast
- — Cheap
- — Template-based
- — No human touch
Good for fun. Not for the real thing.
MusAIc by ~MAK
- — Personal
- — Produced
- — Story-driven
- — Human-guided AI
Built from your story. Produced by a human. Delivered on time.
Every commission includes
- Every version we make, yours to keep
- Naming and title craft
- Artwork for your song page
- Licensing certificate
Out in the world (optional)
- Mastering, $25 for each take you pick. Put that song out later and its $25 comes off the price of Out in the world, because that already includes the mastering — you never pay for it twice.
- Out in the world, $150 — mastered, artwork, metadata, the AI disclosure, delivery to Spotify, Apple Music, YouTube, Amazon and the rest, the store here, Content ID protection, and the splits set up. It earns fifty-fifty, agreed before a word was written.
Choosing your sound
Point at a sound. Heard something on our roster that feels right — Section Line's pedal-steel warmth, Reed Township's boots-on defiance? Name it. Your song won't be by that band and won't copy it; it'll be built in that climate, as its own thing. (Any song by anyone works as a reference too.)
Point at a feeling. "Something warm and acoustic" or "something that starts quiet and builds" or "90s R&B but modern." We translate moods into production choices.
Producer's choice. Trust ~MAK. Give us the story and let the Song Architecture decide what it sounds like. This is often the most surprising result.
The easiest way: point at one of ours
Describing a sound in words is hard. Playing one is easy. There are hundreds of finished songs on this site — put a few on, and if one of them sounds the way you want yours to sound, you have three moves. Any of them is a fine answer.
- Tell us, and we build yours in that climate. Your song won't be that song and won't copy it. It'll be built in the same weather, out of your story.
- Or just buy that one. Sometimes the song that already exists is the song. It's finished, it's mastered, and it's for sale — that's a shorter road and a cheaper one, and we'd rather you took it than paid us to make you something you already found.
- Or start there and go somewhere else. Use it as the jumping-off point and we'll take it in whatever direction your story asks for.
Every song there carries a short code under the title, starting MK-. Send us the code and there's no ambiguity about which one you meant.
Go listenWhat happens when you name a song
Point at a record — ours, or anything you have ever loved — and it goes into a tool we built for exactly this. It doesn't listen to the song. It reads what has been written about that kind of music: criticism, reviews, encyclopedia entries, label and artist copy. At least three real sources, each one named and linkable, because a claim we can't point at is a claim we don't use. Content farms, machine-written encyclopedias, anonymous fan wikis and forums are refused outright — a profile built on those is poisoned at birth.
What comes back isn't a description of that record. It's a climate. Where the tempo sits and which way the key leans. How the dynamics rise and fall across a song. Whether the bass is locked to the kick or walking away from it. What the snare is doing, what the guitars are doing, what the form does. What the production sounds like, what its signature moves are — and, just as usefully, what has to stay out of the mix.
It also captures how that kind of music writes: diction, how long the lines run, what it makes pictures of, how funny or how dark it's willing to be. Two records in the same genre can write completely differently — one all tight swing and chant choruses and plain speech, the other riff-first and inward and heavy. Genre alone would have told us nothing. That's the half most people miss, and it's the half that decides whether your song sounds like it belongs to you.
The artist's name stops there. It stays in an internal note and never travels into the profile, into the production, or onto your record. Not once, not as a wink, not in a nudge at one of their album titles. That isn't squeamishness — it's the same line the rest of this page draws: a style belongs to everyone, a voice belongs to somebody.
And when the tool can't resolve a reference confidently, it says so and hands it to ~MAK rather than guessing. A wrong profile is worse than no profile. Nothing it produces is used on anybody's song until he has read it and signed off. Every reference that does resolve joins the library — so the next person who asks for that climate gets an answer that has already been checked, and the whole thing gets better the more people bring us.
Influences, handled honestly
Tell us 'think [your favorite artist]' and here's what happens: we translate that influence into craft decisions — tempo, delivery, instrumentation, attitude — and build an original in that tradition. What we never do is imitate an identifiable voice; that's someone's identity, it's protected by law, and it's on our refuse list. A style belongs to everyone. A voice belongs to somebody.