MODELS
hexgrad/Kokoro-82M
by hexgrad
Overview
Kokoro-82M is an open-weight text-to-speech model from hexgrad with 82 million parameters.
Details
Kokoro-82M is a hexgrad text-to-speech model listed on Hugging Face. Its model card describes Kokoro as an open-weight TTS model with 82 million parameters, Apache-licensed weights, GitHub and demo links, and a v1.0 release published on 2025-01-27. The official hexgrad/kokoro GitHub repository describes itself as an inference library for Kokoro-82M and links back to the Hugging Face model page and PyPI package.
When to Use
Use Kokoro-82M when you need an open-weight text-to-speech model to evaluate or integrate. Use the official hexgrad/kokoro repository when you want an inference library connected to the Hugging Face model. Use it for research or prototyping where the supplied Apache-licensed weights are acceptable.
Getting Started
- Open the Hugging Face model page for hexgrad/Kokoro-82M.
- Read the model card README for release
- license
- demo
- and usage context.
- Review the official hexgrad/kokoro GitHub repository for inference-library information.
- Run a small text-to-speech trial before relying on it in production.
Key Features
- •Text-to-speech model
- •82 million parameters
- •Open-weight model
- •Apache-licensed weights
- •Official GitHub inference library available
Capabilities
- •text-to-speech
- •speech-synthesis
- •model-inference
Last updated Jun 4, 2026