Kokoro

// Overview

Kokoro is an open-weight text-to-speech (TTS) model with 82 million parameters, paired with the `kokoro` Python inference library. Despite the small size, its authors say it produces quality comparable to larger models while running faster and at lower cost. The weights are Apache-licensed, so you can use it in both production and personal projects.

It's aimed at developers who need to turn text into speech without standing up heavy infrastructure. Because the model is small, it can run on modest hardware, including CPUs, which makes it a fit for local apps, batch jobs, and experiments where a large hosted TTS service would be overkill.

Within the speech and audio space, Kokoro covers the core text-to-speech task across several languages, including American and British English, Spanish, French, Hindi, Italian, Japanese, Brazilian Portuguese, and Mandarin Chinese. Under the hood it uses the `misaki` grapheme-to-phoneme library and `espeak-ng` for some fallbacks and non-English languages.

// What it does

• Compact 82M-parameter model that runs on modest hardware, including CPUs
• Apache-licensed open weights, usable in production or personal projects
• Simple Python API through the `KPipeline` class
• Multilingual via per-language `lang_code` values (American/British English, Spanish, French, Hindi, Italian, Japanese, Brazilian Portuguese, Mandarin Chinese)
• Selectable voices (for example `af_heart`) plus adjustable speed and text-splitting
• Outputs 24 kHz audio you can save directly with soundfile

// Getting started

Install the kokoro package and espeak-ng, then create a pipeline and generate audio from text. The example below mirrors the README's basic quickstart.

pip install kokoro>=0.9.4 soundfile
apt-get -y install espeak-ng

from kokoro import KPipeline
import soundfile as sf

pipeline = KPipeline(lang_code='a')
text = 'Kokoro is an open-weight TTS model with 82 million parameters.'
generator = pipeline(text, voice='af_heart')
for i, (gs, ps, audio) in enumerate(generator):
    print(i, gs, ps)
    sf.write(f'{i}.wav', audio, 24000)

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

// When to use it

• Add local text-to-speech to an app without depending on a hosted TTS API
• Generate voiceovers or audio versions of articles and documents in batch
• Run speech synthesis on CPU-only or low-resource machines
• Prototype multilingual voice features across English, Spanish, French, Hindi, and more

// How Kokoro compares

Kokoro alongside other open-source audio, music & voice tools AI/TLDR tracks, ranked by GitHub stars.