How we make records
The words are ours and a person edits every line. The voices are synthetic. The artists are characters. Here is exactly what that means.
Every record Nity releases is made the same way, and we would rather you heard it from us than found it in a metadata field.
The words are ours, and a person is accountable for every line
A person writes the brief: what the song argues, the constraint it has to survive, the thing that cannot be said. A person writes the genre definition and the artist the song is for. Drafting is machine-assisted from that material, and then a person edits every line, rewriting rather than approving.
No model is asked to invent what a song is about. That is the part that decides whether a song is any good, and it is the part we do not hand over.
Where a client's own language survives into the finished lyric, the person who wrote it is credited as a co-writer. A company cannot be a songwriter; a person can.
The voices are synthetic
The performances are generated rather than recorded in a room, and they carry the provenance watermark that comes with that. When a distributor asks whether a release includes AI-generated music or vocals, we answer yes, because it does.
The artists are characters
Our acts are invented. They are not people with stage names, and none of them is modelled on a real performer. Each one is defined in writing before anything is made: a voice, a face, a way of standing, a set of things they would never do. The same definition produces the record and the press shot, which is why the person on the cover is the person in the photographs.
We do not imitate living artists. A voice is specified in the terms a singing teacher would use, never by naming somebody to sound like.
Why we say so
Three reasons, in the order they matter.
It is going to be visible anyway. The disclosure travels with the release through the distributor, and the provenance is in the audio. Being found out later is worse than saying it first.
It is the interesting part. Building something that takes a brief to a finished master is a more interesting thing to have done than pretending it happened in a studio.
And the alternative is a lie of omission, which is a poor way to open a relationship with somebody about to give you three minutes of their attention.
What we will not do
Clone a real person's voice. Pass a synthetic performance off as a human one. Mass-produce tracks to game a recommendation system. Claim a person wrote what a person did not write, on a release or on a registration.
What we will not say
How any of it is done. The tools, the specifications and the process are ours, the same way a producer's chain and a mixing engineer's moves are theirs. What you are owed is an honest account of what a record is, which is above. The how is the work.