A studio-quality clone of your voice
Make your own singing voice clone in minutes. Turn any vocal into your voice, or create new vocals from scratch.
Privacy and data
Your voice, your data—private.
A voice model is as personal as a recording of yourself. Three things govern who can reach yours, and where the voices we host came from.
Private by default
Every model you train starts closed. No other user can hear it, use it, or find it on the platform — there is nothing to switch off, because nothing is open.
Shared only by email, and revocable
Access opens one person at a time, to the email address you enter — they get the voice and nothing else. Withdraw it whenever you like and their access closes.
Never scraped from the web
Kits' voice and instrument models accessible via the model select modal, have been Fairly Trained certified. No voice in the Kits library was taken off the internet.
Fairly Trained certifiedMore ways to use your voice.
Put your voice on any song
Turn an existing vocal into your voice. Hear yourself on a new song before recording a take.
Collaborate without another session
Use a collaborator’s shared clone, with their permission, for demos, revisions, and harmonies.
Generate vocals in your voice
Create new vocals with the Lead Vocal Generator, then make them yours with your clone.
Upgrade your rough takes
Turn phone recordings and home takes into cleaner vocals using your professional voice clone.
Two ways to clone
Your Vocal Partner For Any Process
Train a studio-grade model of your own voice, or hear a rough one in minutes. Both support singing and speaking, and both output dry vocals you can take straight into a mix.
Studio-grade
Professional Voice Cloning
A voice model trained on a high volume of your own recordings. The one you build once and use across a record — the clone that has to hold up when the vocal sits exposed in the mix.
- Audio needed
- 15+ minutes
- Training time
- ~30 minutes
- Output
- Dry vocals
Best when your source recordings are clean, dry and consistent across the set.
Start a professional cloneFast on-ramp
Instant Voice Cloning
A rough clone from a single reference file. Built for the moment you want to hear whether the idea works, before committing to training a full model.
- Audio needed
- 15+ seconds
- Training time
- Minutes
- Output
- Dry vocals
Rougher than a professional model, particularly on voice likeness.
Start An Instant CloneGetting a good clone
Handheld guidance for the best clone possible
Output quality follows input quality — so we don’t leave you to work that out on your own.
Tutorials and guidance, built in
Walkthroughs for every step of training a voice model, so the first clone you make is one you’d actually use on a track.
Upload, or record with us
Bring vocals from your machine, or follow our guided recording process and we’ll take you through it take by take.
Questions
The things producers actually ask
A clone of your own voice is built from audio you own, and the output is yours to release — the same as any other recording of yourself. What you need rights to is the underlying song and any source audio you convert.
Two things worth keeping straight. Outputs made with Kits Library Voices are 100% royalty-free and cleared for commercial release, with no attribution and no payout owed, provided you own or have rights to the audio being converted. That clearance is specific to library voices and does not extend to cloned voices — a clone carries whatever rights its source audio carries.