Local text-to-speech • 28 voices • Browser-based AI • Privacy-first

Text to Speech

Convert text to natural-sounding speech locally in your browser. 28 built-in voices, adjustable speed, privacy-first — no server upload needed.

What this text-to-speech tool does

This tool uses the Kokoro-82M AI model to convert text into natural-sounding speech directly in your browser. It supports 28 built-in voices across American and British accents, adjustable speech speed, and long text with automatic segmentation. All processing happens locally — your text never leaves your device.

How to use text-to-speech in the browser

  1. 1.Type or paste your text into the input area
  2. 2.Choose a voice and click to preview how it sounds
  3. 3.Adjust the speech speed if needed
  4. 4.Click Generate and download the audio

Why this text-to-speech workflow is useful

All processing happens in your browser after model loading

28 built-in voices with American and British accents

Adjustable speech speed from 0.5x to 2.0x

Supports long text up to 5000 characters with automatic segmentation

Text to Speech FAQ

Is my text kept private?

Yes. Text-to-speech processing runs entirely in your browser. Your text is never uploaded to any server — everything stays on your device.

How many voices are available?

There are 28 built-in voices including 11 American female, 9 American male, 4 British female, and 4 British male voices. Click any voice to preview it before generating.

Why is the first generation slower?

The first time you generate speech, the AI model (~154MB) needs to download to your browser. This is a one-time download — the model is cached and reused on subsequent visits. A WebGPU-capable browser is required (Chrome 113+ or Edge 113+).

What are the system requirements?

You need a WebGPU-capable browser such as Chrome 113 or later, or Edge 113 or later. The tool downloads a ~154 MB AI model on first use, which is cached in your browser for future visits. Processing speed depends on your GPU — faster GPUs generate speech more quickly.

Will this slow down my computer?

Running the AI model consumes significant GPU and memory resources. On low-end or older devices, you may experience browser lag, high memory usage, or temporary unresponsiveness during processing. For the best experience, close unnecessary tabs and applications, and use a device with a dedicated GPU if possible.

How natural does the speech sound?

The Kokoro-82M model produces natural-sounding speech with distinct voice personalities. Quality varies by voice — some sound more conversational while others are more formal. You can preview any voice before generating to find the one that fits your needs.