Kokoro 82M with zero data retention.

A light open-weight voice model, at a fraction of a cent. On Thinkr, it runs with zero data retention (ZDR), so the provider keeps nothing.

›open/kokoro-82m

Maker
hexgrad
Feature
Audio · Voice
ZDR
$0.001 per 1,000 characters

How private it is

Most voices run with zero data retention: the provider reads your script to speak it and keeps nothing. The music of Lyria 3 and the voices of MiniMax and Grok run in Open, where the provider may keep a request for a while. Your audio stays in your browser, out of any backup.

Make your first sound now.

Make audio