MelogniteNo neural network.
No generative model.
Not one song stored inside.
Generative AI writes music by predicting what millions of scraped recordings make statistically likely. Its knowledge is billions of unreadable parameters — large enough, as researchers keep demonstrating, to memorize and leak fragments of the very songs it was fed.
Melognite is built the other way around. Harmony, cadence, voice leading and song form are written down as explicit rules — drawn from published music theory, music psychology and genre analysis, then implemented as code you can inspect. Every choice follows from a musical principle — how tension is built and released, how a line resolves, how a phrase closes — not from imitating a recording. Open the style packs and look for a melody: there is none — and unlike a neural network, you can trace every note back to a reason.
That's why every note has a reason you can trace: a rule, a context, a choice. Deterministic — the same seed with the same settings yields the same piece, bit for bit, verifiably. And it has no song library to copy from: it cannot reproduce a recording it was trained on, because it was never trained on any.
Same musical conclusions as any trained model, sometimes — a V–I cadence is likely either way. Fundamentally different custody of other people's work.
How it actually works
Isn't music theory itself just statistics drawn from existing songs — exactly what AI does?
Not the way an AI model is. Music theory is what centuries of musicians distilled from music: tested, taught, corrected, written down as principles, with the actual songs deliberately left behind. A dictionary is made by reading millions of texts, yet it contains words and grammar, not books — nobody accuses a dictionary of plagiarizing novels. A generative model is the opposite: it keeps the surface. Its billions of parameters are demonstrably large enough to memorize fragments of specific recordings. Melognite ships its musical knowledge in readable files you can open: 619 chord progressions stored as scale degrees (the longest is eight steps) and 902 rhythm figures stored as strike patterns without pitch. The pitch is computed while the song plays, from the chord underneath. Every number in a pack is a range, a weight or a proportion — never a tune. Same musical conclusions, sometimes. Fundamentally different custody of other people's work.
If it follows fixed rules, why doesn't everything sound the same?
Rules define a space, not a point. The rules of chess are fixed and finite; no two games are alike. Within everything the theory allows, Melognite makes seeded choices — which motif, which cadence approach, which rhythm. Change the seed, you get a different valid piece. Keep the seed and your settings, you get the identical piece, forever.
What exactly is inside Melognite, if there's no trained model?
Two things, both inspectable. One: rules as code — scales, functional harmony, cadences, song forms. Two: a vocabulary of abstractions — chord progressions as scale degrees and rhythm figures as strike patterns without pitch; the pitch is computed at play time from the chord underneath. No melodies. No recordings. No weights. No black box.
Could Melognite accidentally reproduce an existing song?
Not the way generative models have been shown to. Melognite has no song library to copy from, so it cannot reproduce a recording it was trained on — it was never trained on any. Like any newly written music, an accidental resemblance to one of the millions of existing melodies can never be ruled out entirely — the same is true for every human composer. When it happens, it is the resemblance of shared grammar (two writers using a V–I cadence), not of a shared source.
What does "deterministic" mean here — and why should I care?
Same seed and settings, same song, bit for bit — verifiable by checksum. Your instrument, style and slider choices are part of the recipe, not of the seed — a share link (Basic and up) carries both. It matters for ownership and for editing: because every note follows from a rule you can see, you change the decision instead of re-rolling a slot machine. Generative AI offers neither; its output is a sample from a distribution even its makers can't fully explain.
So is it AI or not?
Depends on the dictionary. Computer science has called rule-based composition "algorithmic" since the 1950s — long before anything scraped the internet. What "AI" means today — networks trained on other people's work, producing outputs nobody can explain — is precisely what Melognite is not. No training run, no dataset of recordings, no network. Every rule was written down on purpose, by a person, and you can go read it.
Who decides whether it actually sounds good?
Musicians — not metrics. Rules can guarantee that a progression is valid and that a phrase closes properly; only a trained ear can tell whether a chorus actually lands. So before launch, every genre goes through extensive listening tests with experienced musicians, and each one is fine-tuned in close collaboration with them — genre by genre, style by style.
The difference isn't that we reach different conclusions — it's that with Melognite you can read every conclusion, and no one else's song is stored inside — open the packs and check.