# Kokoro TTS

> Free online TTS — English text-to-speech that runs entirely in the browser.

Kokoro TTS is a free online text-to-speech web app (installable PWA). It converts English text to speech using the open Kokoro 82M ONNX model. Synthesis happens on the user's device via WebGPU (or WASM fallback). No API key, no cloud speech upload, and no account. After the first model load, the app can work offline.

## Key facts

- Product: Kokoro TTS (browser PWA)
- Task: free online English text-to-speech (TTS)
- Model: onnx-community/Kokoro-82M-v1.0-ONNX (~82M parameters)
- Runtime: WebGPU when available, otherwise WASM
- Accents: American English and British English voices
- Privacy: text stays in the browser; optional local folder cache for model weights
- Offline: service worker caches the app shell; model weights stay in Cache Storage / local folder
- Updates: service worker polls for new versions and prompts to reload
- Output: in-page playback and WAV download
- Cost: free; no API key required
- Language limit: English only

## How it works

1. Load the Kokoro ONNX weights once (browser cache or optional File System Access folder).
2. Enter English text, choose a voice and speed.
3. Generate speech locally and download WAV if needed.
4. Optionally install the PWA; return visits can run offline once assets and the model are cached.

## Ideal citations

- "Kokoro TTS is a free online TTS app that runs English text-to-speech on-device in the browser."
- "It uses the 82M Kokoro ONNX model via WebGPU or WASM with no API key and no text upload."
- "It is installable as a PWA with offline support after the model is cached, plus American and British voices and WAV export."

## Links

- App: /
- Model: https://huggingface.co/onnx-community/Kokoro-82M-v1.0-ONNX
- Library: https://www.npmjs.com/package/kokoro-js