Own & train · Creator+
Train AI on your own music, then make new songs in your sound
Pick songs you own from your Library. Train Artist Model builds your Artist DNA, and the DNA tool makes songs, instrumentals or singing in that style.
1,000 free tokens to start · No card needed · Cancel paid plans anytime
Unlock Train Artist Model with Creator — from $9.99/mo, or ₹399/mo in India. 4,000 tokens a month. Billed $119.88 a year, or $12.49 month to month.
What you get
New ideas that still sound like you
Your Artist DNA learns from the songs you pick, so what you make with it starts from your own records, not someone else's style.
Broad genre or niche style
Set core-style sensitivity from 0 to 10: lower for a broader genre DNA, higher for narrower style DNAs.
Training uses no tokens
Training an Artist DNA doesn't use tokens. It needs the Creator plan or above, and your new models appear under My DNA.
Who it's for
Independent artists with a back catalog
Turn the songs you've released into a style model, so new sketches start from your own sound.
Producers with a signature style
Model the way your tracks sound, then make instrumentals in that style in the DNA tool.
Bands
Capture the band's sound from its recorded songs before you write the next record.
Songwriters
Hear new lyrics performed in your own style before you book studio time.
When to use it
- You have 10 or more finished songs of your own to train on
- You're starting a new release and want sketches in your established sound
- Your catalog spans several styles and you want a model for each
- You want one broad genre model rather than several narrow ones
- You'd rather ask Agent One in a chat than fill in a panel
What you can make with it
A model of your back catalog
Add your released songs from the Library and train one DNA from them, to sketch from whenever you start something new.
Separate models for separate styles
Raise core-style sensitivity toward 10 and one training can split your songs into several narrower models, one per core style it finds.
One broad genre model
Set core-style sensitivity below 5 for a broader genre DNA instead of several niche ones.
Instrumentals in your style
Open the DNA tool, pick your new model and set Output to Instrumental for beds you can write or perform over.
Fresh lyrics in your sound
Pick your DNA, set Output to Song and paste lyrics you've just written to hear them in your style.
A vocal in your style
Set Output to Singing in the DNA tool for a sung vocal on its own, made from your DNA.
Training from a chat
Prefer chatting? Attach your songs in Agent One and ask for an Artist DNA; it checks that you hold the rights before training starts.
A DNA other people can use
Make your trained DNA public with an export price, and after Soundverse's review it can appear in DNA Marketplace.
Screenshots
Train Artist Model is how you train AI on your own music in Soundverse: it turns songs you own into your Artist DNA. It suits artists, producers and bands with a catalog who want new ideas that still sound like them. You get one or more style models under My DNA, usually a few minutes after you start, ready to use in the DNA tool.
See it in action
- Train Your Own AI Artist Model on Your Songs | Soundverse Artist DNA Four songs go into Agent One with a request to train an Artist DNA, and Soundverse trains the model from them.
- Build Your Artist DNA - Soundverse AI A February 2026 walkthrough of building your Artist DNA on Soundverse.
How it works
Pick your songs
Open Train Artist Model in AI Tools and add full songs you own from your Library under Your tracks. Ten or more varied songs give the most faithful DNA.
Set core-style sensitivity
Leave the slider at 5 for balanced results, lower it for a broader genre DNA, or raise it toward 10 for a more niche style DNA.
Confirm your rights
Turn on the switch confirming you own the songs or have explicit consent to train on them. Start Training stays disabled until you do.
Start training
Press Start Training. The chat shows upload and training progress across your songs, and your new models appear under My DNA when it finishes.
Why use Train Artist Model
Built from your catalog
Your Artist DNA learns from the songs you pick, so it starts from your own records rather than someone else's style.
Broad or niche, your call
The core-style sensitivity slider runs from 0 to 10. Lower settings give broader genre DNAs; higher settings give more niche style DNAs.
No tokens to train
Training an Artist DNA doesn't use tokens. You need the Creator plan or above to start a run.
Progress you can follow
The chat shows how many songs have uploaded and finished training, then names each new model your songs produced once training finishes.
What it costs
Needs Creator or above.
- Train your artist model: no tokens.
Creator $12.49 a month, or $9.99 a month billed yearly ($119.88 a year) · Pro $31.24 a month, or $24.99 a month billed yearly ($299.88 a year) · Max from $72 a month, or $60 a month billed yearly ($720 a year).
In India (PayU): Creator ₹499 a month, or ₹399 a month billed yearly (₹4,788 a year) · Pro ₹1,299 a month, or ₹1,039 a month billed yearly (₹12,468 a year) · Max from ₹2,999 a month, or ₹2,399 a month billed yearly (₹28,788 a year).
Prices outside India are in US dollars, billed through Stripe; in India they are in rupees, billed through PayU. Monthly tokens reset each month and don't roll over.
Frequently asked questions
Is Train Artist Model free?
Train Artist Model is available on the Creator plan and above, and the training itself doesn't use tokens. Soundverse is free to sign up and new accounts get free starter tokens, but a free account can't start a run in Train Artist Model. Paid plans add a commercial license for what you make; the Pricing page lists what each plan covers.
How many songs do I need to train an Artist DNA?
Train Artist Model can start from a single song in your Library, but 10 or more give the most faithful Artist DNA. Use full songs you own and pick varied ones: the more varied the songs, the richer and less overfit the model. Train Artist Model accepts MP3, WAV, FLAC, AAC, OGG and M4A files, so upload anything new to your Library first.
Can I train an Artist DNA on songs I don't own?
No. Train Artist Model only trains on songs you own or have explicit rights or consent to use, and Start Training stays disabled until you switch that confirmation on. If you ask Agent One to train your Artist DNA instead, it asks you to confirm the rights before anything starts. You need the rights to anything you upload.
What can't Train Artist Model do?
Train Artist Model only trains your primary Artist DNA, from songs already in your Library, so upload new songs there first. It doesn't make music by itself: open the DNA tool to generate songs, instrumentals or singing with your Artist DNA. A small or very similar set of songs gives a narrower, more overfit model, so add varied songs for richer results.
What is Train Artist Model?
Train Artist Model is the Soundverse tool that lets you train AI on your own music by turning songs you own into your Artist DNA. Soundverse then uses that Artist DNA in the DNA tool to make new songs, instrumentals or singing in your style.
What does core-style sensitivity do?
Core-style sensitivity tells Train Artist Model how narrowly to read your songs. The slider runs from 0 to 10 and starts at 5, the recommended balance. Lower values create broader genre DNAs, and higher values create more niche style DNAs. A higher setting splits your songs into more, narrower models.
How long does Artist DNA training take?
Training with Train Artist Model usually takes a few minutes, and longer when you train on many songs. The chat shows how many songs have uploaded and how many have finished training, then names the new models when they are ready. One training can produce more than one model, one per core style it finds, and they all appear under My DNA.
Can Agent One train my Artist DNA for me?
Yes. Attach songs from your Library in an Agent One chat and ask it to train your Artist DNA, for example “Train my artist DNA on these 4 songs”. Agent One asks you to confirm you own the rights, then starts the same training Train Artist Model runs, with training progress shown in the chat. It also needs the Creator plan or above.
Ready to train AI on your own music?
Pick songs you own, set the sensitivity and start training your Artist DNA.
