Opening audio tool…
Text to speech in your browser
Generate spoken audio from text with on-device Kokoro text to speech. Choose an English voice and speed, preview the result and download a WAV file.
How to use text to speech
- Enter or paste up to 12,000 characters of text.
- Explicitly load the Kokoro model, then choose a voice and speed.
- Generate the speech, listen to the preview and download the WAV.
Before you start
Four English voices are available. The first run downloads model files from Hugging Face and a runtime from jsDelivr. Generation uses your device's memory and WebAssembly rather than a server inference service.
Audio models load only when you request them. Files and text are not sent to an AI inference service. Read the privacy policy.
Related audio tools
Browse all audio tools for format choices, file limits and common questions.