András Ferencz
Text-to-speech model

Kokoro

Kokoro-82M, an open-weight text-to-speech model with 82 million parameters, small enough to run on a laptop.

View on GitHub
How I use it

A small, open text-to-speech model that runs locally.

What it is
Text-to-speech model
On my tools page
Video and voice
Repository
hexgrad/kokoro
Licence
Apache-2.0
Updated
November 2026