Guide
English
Search the docs⌘K
5.4

Voice cloning

Read a given passage out loud and you get a voice of your own in DOTV — from then on it sits alongside the 320 system voices, ready to read anything you pick it for, so the narration and the characters' lines in your AI video can be in your own voice. This section covers how to clone one, a few rules, and how to manage your voices.

How to clone one

The way in is the record button in the top right of the voice picker, which opens Clone a voice:

  1. The panel gives you a passage under Read this: (three of them to swap between, about 20 seconds each). Swapping gets you a different set of phonemes, and saves you reading the same line every time.
  2. Hit record, read it through, hit it again to stop. You can play the recording back, and hit Record again if you don't like it.
  3. Tick the usage rules on the line below (open the Voice Cloning Rules to read them in full).
  4. Submit. The voice you get shows up on the My voices page.
Clone a voiceRead the script under Read this: out loud (there are three to swap between); tap the pill to start recording, tap again to stop. Play it back or record again afterwards. You have to accept the usage terms before you can submit.
The passage for this round on top, the pill in the middle that starts recording, the usage rules at the bottom.

A few rules

  • You may only clone your own voice, or one you have explicit permission to use.
  • It must not be used to fake an identity, impersonate anyone, defraud, or for anything else unlawful.
  • The voices you create are stored under your account, so they're still there when you sign in on another machine.

Managing your voices

On the My voices page of the voice picker:

  • Rename — double-click the name on that row and edit it in place. The default is "My voices N"; something like "Wang — narration" is easier to find.
  • Delete — Remove voice, under More on that row. What goes is this entry in your own list.
  • Cloned voices can be favorited and previewed just like system ones (the preview is this voice reading the same greeting the system voices read, so you can compare them side by side).
  • They're filed under Chinese (Mandarin), so they sit with the Chinese voices in the language filter.
Try it in DOTVReading only goes so far: open DOTV and make an AI video with the feature this section covers.