# kokoro-web

> Kokoro Web is a self-hosted, open-source AI voice generator with OpenAI-compatible API and browser UI for natural speech in multiple languages.

- **Category:** Text to Speech
- **Licence:** MIT
- **Latest version:** v0.1.3
- **Open source:** oss
- **Languages:** Typescript, Svelte
- **Platforms:** Docker Compose
- **Tags:** Machine Learning, Multilingual, Docker, AI, Text To Speech, Voice, Self Hosted, Text To Speech, Webgpu, Openai compatible, Voice generation, Kokoro tts, Kokoro, Kokoro 82m

## Description

Kokoro Web is a free, open-source AI text-to-speech tool that runs in the browser or self-hosted via Docker. It uses the Kokoro-82M model from Hugging Face to produce natural voices across English (US/UK), Japanese, Chinese, Spanish, Hindi, Italian, and Portuguese (Brazil). The self-hosted edition exposes an OpenAI-compatible endpoint at /api/v1, letting existing apps swap in this service without code changes. The web interface offers dozens of voice profiles, speed adjustment, synthetic pause tags, WebGPU acceleration, and multiple model quantization levels. Optional API key authentication is controlled by an environment variable. The project is MIT licensed and maintained by Eduardo Lat.

## Links

- [Homepage](https://github.com/eduardolat/kokoro-web)
- [Website](https://voice-generator.pages.dev)
- [Docker](ghcr.io/eduardolat/kokoro-web)
- [Demo](https://voice-generator.pages.dev)
- [GitHub Repository](https://github.com/eduardolat/kokoro-web)
- [API Documentation](https://voice-generator.pages.dev/api/v1/index.html)
- [Author Website](https://eduardo.lat)

---
Source: https://selfhost.directory/project/kokoro-web