Kokoro-82M

Kokoro-82M is a compact open text-to-speech model with 82 million parameters, built on StyleTTS 2 and ISTFTNet. Per the model card, version 1.0 (published 2025-01-27) has 54 voices in 8 languages: American and British English, Japanese, Mandarin Chinese, Spanish, French, Hindi, Italian and Brazilian Portuguese. It has no voice cloning: you choose from the ready-made voices. The author notes that support for non-English languages may be thin, with only a few voices in some of them. The authors state that it was trained on permissive and synthetic audio only. The card does not state a VRAM requirement; the small size makes it easy to run, with a pip package and the espeak-ng tool. It is the most downloaded text-to-speech model on Hugging Face, although it is the oldest of the models we list. The weights are public without any acceptance step under the Apache-2.0 license, which allows commercial use. The author offers an official demo Space on Hugging Face, and the code is on GitHub. Date of check: 2026-10-19.