Read it! – Free Local Text to Speech Voice Reader

Read it! turns any web page into audio. Select some text and click Read, or press play to hear the whole article. Natural-sounding AI voices read it to you, and the sentence being read is highlighted so you can follow along. Everything runs on your own computer. The voice model works inside your browser, so the text you listen to is never sent to a server. There is no account, no subscription and no usage limit. WHY PEOPLE USE IT • Listen to long articles, documentation or newsletters while you walk, cook or rest your eyes • Understand dense text better by hearing it and seeing it at the same time • Proofread your own writing by listening to it • Support for dyslexia, low vision or reading fatigue FEATURES • 50+ natural voices in English (US and UK), Spanish, Italian, Portuguese, Hindi, Japanese and Chinese • Sentence highlighting that follows the voice and keeps the current sentence in view • Click any sentence to jump there • Speed from 0.6x to 2x • Finds the main article and skips menus, ads, cookie banners and comments • Reads image descriptions, skips code blocks and struck-out text (each can be turned off) • Reads quotes in a second voice • Switches to a matching voice when a page marks text in another language • Says numbers, dates, prices, units and ranges the way people do: "€50" becomes "50 euros", "2024-01-15" becomes "January 15th, 2024", "70%-80%" becomes "70 to 80 percent" • Skips emojis if you prefer, and lets you add your own pronunciation rules • Remembers settings per site, including which part of the page to read • Keeps the screen awake while reading • Light and dark theme KEYBOARD SHORTCUTS • Alt+R: read the selection or the page • Space: play or pause • Left / Right arrows: previous or next sentence • Up / Down arrows: faster or slower • Esc: stop PRIVATE BY DESIGN • Speech is generated locally with the open Kokoro model, running on WebGPU • Your pages and selections stay in your browser • No analytics, no tracking, no sign-up • Settings are stored with Chrome's own storage GOOD TO KNOW • The first time you press play, the voice model (about 330 MB) is downloaded once from Hugging Face and stored in your browser. After that it starts quickly and works offline. • A recent computer helps. A graphics card with WebGPU support makes speech generation much faster. OPEN SOURCE Read it! is free and open source (MIT). Speech comes from the Kokoro-82M model (Apache 2.0) and eSpeak NG (GPL-3.0). Full credits are included in the extension. Source code: https://github.com/hakubo/ReadIt