dub44Sign in

Hear yourself say
something you never said.

20 seconds of your voice, then type anything. No account, no card — and your recording is deleted the moment the clip is made.

Hear it first — 15 seconds

One voice, twice. The first clip is a real recording; the second is that same voice reading a sentence it never said.

What was recorded

0:00 / 0:00

The same voice, reading new text

0:00 / 0:00
  1. 1
  2. 2
  3. 3

Step one

Talk for 20 seconds

Don't read anything — just answer this out loud: What are you working on at the moment, and what is annoying about it?

Already have twenty seconds of your voice on your phone? Use that.

Step two

Type what you never said

62 / 140Record or upload first

What happens to your recording

  • It is used once, to make your clip, and deleted in the same breath. It is never stored, never used to train anything, and never leaves the EU.
  • Every clip carries an inaudible provenance watermark — the free one and the paid ones alike.
  • Only clone a voice you have permission to use. Here that means your own.