Create · Creator+
Sing new songs in your own voice from one short recording
Pick a take with 15 seconds or more of singing, confirm it's your voice, and train a clone the AI Song Generator can sing in. On the Creator plan and above.
1,000 free tokens to start · No card needed · Cancel paid plans anytime
Unlock Train Voice Clone 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
Demos in your voice, no session
Turn on Use custom voice in the AI Song Generator, pick your clone, and new songs are sung in it.
One take is enough
A single recording with at least 15 seconds of singing becomes a voice you can reuse for every new song.
Consent built in
Training can't start until you confirm the voice is yours or that the singer has given explicit consent.
Who it's for
Singers
Hear new songs in your own voice without recording every take.
Songwriters who sing their demos
Make pitch demos in your voice from new lyrics in AI Song Generator.
Independent artists
Use one lead voice across every song in a project, even when the songs are made weeks apart.
Producers with a consenting vocalist
Write songs in a collaborator's voice once they've given explicit consent.
When to use it
- You want demos in your voice but can't record every idea
- You have a clean, dry recording in which you sing for 15 seconds or longer
- A singer you work with has agreed to let you write in their voice
- You want Agent One to sing new songs in your voice
What you can make with it
Demos in your own voice
Write lyrics, turn on Use custom voice in AI Song Generator and pick your clone to hear the song sung in it.
An EP with one lead voice
Pick the same clone for every track so an EP keeps one lead voice from the first song to the last.
Ask Agent One for it
Ask Agent One for a song sung in your saved voice, and it sings the song with your clone.
A collaborator's voice, with consent
With a singer's explicit consent, clone their voice from a clean take and send them songs to hear before a session.
A clone for each style
Train one clone from a breathy take and another from a belted one, describe each, and pick the one a song needs; each clone is its own training run.
Your voice in a new genre
In AI Song Generator's Custom mode, pair your clone with a Quick Style tag you rarely sing, such as Bossa-Nova or Synthwave, to hear it in that style.
A gift song in your voice
Make a birthday or anniversary song sung in your clone and send it without recording a note.
Screenshots
Train Voice Clone builds an AI voice clone from one short recording, so the AI Song Generator can sing new songs in that voice. It's for singers and songwriters who want demos and finished songs in their own voice without recording every take. You give it a reference with at least 15 seconds of singing, a name and an optional description, and you get a reusable voice listed under My Voices & Models.
See it in action
- Clone Your Voice in One Take, Then Make Songs With It | Soundverse Voice Clone Record one take, ask Agent One to clone it, and use your new voice in the songs you make next.
- How to Clone Your Voice and Make Songs in It with AI | Soundverse AI Voice Clone Clone your voice, then make full songs sung in it without singing every take yourself.
- Make a Song in YOUR Voice with AI: Voice Clone + AI Song Generator | Soundverse AI Record a short sample, create your AI voice, then describe a song and let the AI Song Generator sing it in your voice.
How it works
Pick a reference vocal
Choose a recording from your Library with at least 15 seconds of singing, as an MP3, WAV or FLAC file up to 10 MB. M4A isn't supported yet, so convert it first.
Name and describe the voice
Give the voice a name, and optionally describe its character, for example a warm acoustic vocal with subtle vibrato.
Confirm consent
Tick the box confirming it's your voice or that you have the speaker's explicit consent to clone it. Start Training stays disabled until you do.
Train and reuse
Click Start Training and Agent One creates the clone in your chat. The voice then appears under My Voices & Models, ready for Use custom voice in the AI Song Generator.
Why use Train Voice Clone
One take, a reusable voice
A single recording with at least 15 seconds of singing becomes a voice you can pick again for every new song.
New songs in your voice
Turn on Use custom voice in the AI Song Generator, pick your clone, and the song is sung in that voice.
Consent before training
Training can't start until you confirm the voice is yours or that the speaker has given explicit consent.
All your voices in one list
Every clone you train is listed under My Voices & Models in the panel, so you can find it and use it again later.
What it costs
Needs Creator or above. Cloning with Soundverse v9 needs Pro.
- Train one voice clone: 6,000 tokens (about $14.99–$18.74 on Creator, with a top-up, or ₹599–₹749 in India).
A run costs more than Creator's 4,000 monthly tokens, so on Creator it needs a top-up. Pro includes 10,000 a month.
Agent One runs this tool for you, and the agent's own steps are metered on top of the tool's price: about 7 tokens a step in Lite mode, 54 in Standard mode and 128 in Pro mode, the default on paid plans.
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 Voice Clone free?
No. Train Voice Clone is available on the Creator plan and above, and each training run uses tokens from your plan. Soundverse itself is free to sign up, and new accounts get free starter tokens for the tools on the Free plan. The Pricing page lists what each plan includes.
How many tokens does Train Voice Clone use?
Training a voice clone costs about 6,000 tokens, more than the 4,000 tokens the Creator plan adds each month, so on Creator, expect to add a top-up, or choose Pro, which adds 10,000 a month. You pay once per clone: songs sung in it cost the same as any other song on the model you pick.
How long does my voice sample need to be?
Train Voice Clone needs a reference with at least 15 seconds of singing. The minimum counts the vocal itself, not the file's total length, so a long intro without singing doesn't help. The panel flags a clip shorter than 15 seconds before you submit, though it can't tell how much of it is singing. A clean, dry take that covers a range of notes gives the clone the most to work with.
Can I use songs made with my voice clone commercially?
On the Free plan, what you make on Soundverse is for personal use. Paid plans add a commercial license; the Pricing page lists what each plan covers. Train Voice Clone itself needs the Creator plan or above, and you need the rights to the recording you clone and consent from the person whose voice it is.
What can't Train Voice Clone do?
Train Voice Clone makes a voice for singing new songs in the AI Song Generator on the Soundverse versions; it doesn't re-sing a vocal you already have, which is what Voice Swap does with a model from Train Voice Model. Train Voice Clone needs at least 15 seconds of singing in an MP3, WAV or FLAC file up to 10 MB, and training starts from AI Tools, not inside AI Studio.
Can I clone someone else's voice?
Only with their explicit consent. Train Voice Clone asks you to confirm that the voice is yours or that the speaker has given explicit consent to clone it, and training won't start until you do. Never use a voice clone to impersonate a real person or to make them appear to sing something they didn't agree to.
What is Train Voice Clone, the AI voice clone tool?
Train Voice Clone is a Soundverse tool that builds an AI voice clone from one short recording of a singer. Once the clone is saved, the Soundverse AI Song Generator can sing new songs in that voice whenever you turn on Use custom voice.
How do I make an AI voice clone of my own voice?
Record yourself singing in a quiet room, with a dry vocal and a range of notes and emotions, and add the take to your Library. Open Train Voice Clone, pick the recording, name the voice and tick the box confirming it's your voice. Click Start Training, and when Agent One finishes, your clone appears under My Voices & Models.
What files does Train Voice Clone accept?
Train Voice Clone takes one reference recording from your Library, as an MP3, WAV or FLAC file up to 10 MB. M4A files aren't supported yet, so convert them first, for example with Soundverse's Media Converter. If the recording is on your device, upload it to your Library before you start, and only use recordings you have the rights to.
How is Train Voice Clone different from Train Voice Model?
Train Voice Clone is the fast option: one clip with at least 15 seconds of singing gives you a voice for the AI Song Generator. Train Voice Model needs at least 10 minutes of clean solo recordings and trains a higher-fidelity singing and speaking voice model, which you then use in Voice Swap to re-sing existing vocals.
Ready to sing in your own voice?
Pick a take with at least 15 seconds of singing, confirm it's your voice, and train a clone for every new song.
