Kokoro Web

Kokoro Web is a free, open-source AI voice generator that turns text into speech in the browser. It also supports self-hosted API use with an OpenAI-compatible interface.

AI Speech Synthesis Text to Speech

Visit Website

Overview

Kokoro Web is a free, open-source AI voice generator for turning text into speech in the browser. The site presents it as a self-hostable product with an OpenAI-compatible API for instances that run the model themselves.

The interface lets users choose an execution place, configure API settings for self-hosted setups, pick a language accent and voice, and then generate speech from pasted text. The page also exposes timing controls such as speed and inline pause tags, which makes it suitable for simple narration, testing, and voice-output experimentation.

Features

Browser-based text-to-speech

Generate speech by entering text directly in the browser and selecting a voice before starting synthesis.

Execution placement controls

Choose where generation happens, with a browser option and a self-hosted API option for instances that expose it.

API settings for self-hosted use

Configure an OpenAI-compatible API base URL and API key when using self-hosted API settings.

CPU and WebGPU acceleration

Switch between CPU and WebGPU acceleration, with WebGPU labeled as the faster option.

Model quantization choices

Select from multiple model quantization options with different reported model sizes on the page.

Language and voice selection

Use supported language accents and a long list of voices, with an advanced mode for voice selection.

Use Cases

Quick text-to-speech generation
Create spoken versions of short scripts, drafts, or notes directly in the browser by selecting a voice and generating audio from pasted text.
Voice and accent testing
Try different voices, accents, and speed settings to compare how a line sounds before settling on a final readout.
Timing and pacing experiments
Use inline pause tags and the speed control to shape pacing for narration-style output or pronunciation checks.
Self-hosted deployments
Run the model in a self-hosted setup and point the app at an OpenAI-compatible base URL and API key when you need API-based generation.
API discovery and integration review
Explore the available API endpoints and supported languages or voices through the included Swagger UI documentation.

Pros and Cons

Pros

Free and open source, with the site stating it is free for personal and commercial use.
Supports browser-based text-to-speech generation without requiring a desktop application.
Offers both CPU and WebGPU execution options, with WebGPU marked as faster.
Includes a broad set of voices and multiple language accents on the main interface.
Documents an API through Swagger UI for self-hosted instances.

Cons

The hosted experience appears limited compared with the self-hosted API path, which the site says is required for API access.
The source does not document advanced production features such as project management, team collaboration, or export formats beyond audio generation.

FAQ

How does Kokoro Web work?

Kokoro Web provides an online text-to-speech workflow: enter text, choose a language accent, select a voice, and generate audio in the browser. The source also notes an API for self-hosted instances.

Is the API available in the hosted web app?

The home page says the API is only available for self-hosted instances, and the pricing page mentions an OpenAI-compatible API in that context.

Is Kokoro Web free to use?

The site describes Kokoro Web as 100% free and open source, and the pricing page says it is free for personal and commercial use.

Which languages or accents are supported?

The interface supports multiple language accents, including English (US), English (UK), Japanese, Chinese, Spanish, Hindi, Italian, and Portuguese (Brazil).

Can I control timing in the generated speech?

The page includes a speed control and supports synthetic pauses in text using tags such as [1s] and [0.2s].