Independent artists
No label pipeline and limited studio time: cheap iteration on your own material is the point.
For artists
AI that starts from your catalog, not someone else's
If you write and record your own material, you already have the most valuable training data available to you. ANYANO turns it into a private Sound you can generate from and explore.
Why this matters to artists
The premise
Most AI music tools ask you to describe somebody else's sound. That is an odd request to make of an artist who has spent years developing their own.
Your catalog encodes decisions you no longer have to think about: how you voice chords, where you place a snare, how long you let a section breathe. A model trained on that material inherits those tendencies, which makes its suggestions relevant instead of generically competent.
It also keeps the ethical position simple. You are training on work you made and hold the rights to, which is a very different proposition from prompting a shared model to imitate a named artist.
Workflow
Where it fits in a writing process
- 01
When a demo stalls
Generate around the material you already have to hear plausible continuations, then write the one you actually want.
- 02
When you need volume of options
Sketch several directions for the same idea in an evening instead of committing to the first arrangement that works.
- 03
When you want to hear your older work moved forward
Train on an earlier era of your catalog and generate against it to hear what that identity sounds like pushed somewhere new.
- 04
When you are preparing to record
Use generated sketches to make arrangement decisions before studio time, not during it.
Fit
By kind of artist
Songwriters
Train on your demo archive and generate continuations while you write, with lyrics kept alongside in the app.
Singer-songwriters
Train on instrumental beds or stems if you would rather keep vocal material out of the dataset.
Bands
Train on rehearsal recordings and stems to sketch arrangements between rehearsals.
Composers
Explore variations of a motif across your own harmonic vocabulary rather than a stock library's.
Electronic producers
Stems, loops and sound design in your catalog make unusually good training material.
These are the same product used with different material and intent, which is deliberate — we would rather explain the fit honestly than publish a separate page per keyword. Producers get their own page because the workflow genuinely differs.
Caution
Keeping your identity intact
A personalized model can flatten you if you let it. Generated output tends toward the centre of what you have already done, which is useful for continuity and useless for change. Use it for options and momentum, and keep deliberately making the choices it would not have proposed.
FAQ
Frequently asked questions
Related
Keep reading
Your catalog is training data only you have.
Train a Sound from it and start exploring.