The microphone records only while the button is lit, and the sound is turned into text here, in your browser.
A voice here is a blend of Kokoro's voices. Mix them and you get a voice that belongs to no one.
How the voice is made
- 1 The text becomes phonemes (espeak-ng)
- 2 The phonemes and the blended style go into Kokoro
- 3 Kokoro returns the sound wave, 24 000 samples a second
What the avatar keeps about you from this conversation. Delete a line and it forgets it. Nothing is stored after you close the page.
Nothing yet. Tell it something about yourself.
This shows the idea. In the Mnemosyne OS app, memory lives in vaults on your computer, and you decide what enters, what moves and what leaves.