What it does
You've got a 40-page spec open, or a Wikipedia rabbit hole you've been staring at for an hour. Voice Reader lets you select any text on any page, right-click → Voice Read, and just listen. The whole thing runs offline — no text leaves your machine, no account to create, no subscription to forget to cancel.
It's a free Chrome extension built on Piper VITS, a neural voice model that runs in your browser via WebAssembly. The voice model downloads once (~63 MB) and stays local after that.
Completely Free
No tiers, no trial periods, no "upgrade to unlock speed control." Open source, full stop.
100% Private
Everything runs offline. Your text never touches a server — not even for analytics. There's nothing to opt out of.
Neural Quality
Piper VITS via WebAssembly. It sounds like a person, not a phone tree. Runs directly in the browser tab.
Accessible
Word highlighting follows along in real time, which helps a lot if you have dyslexia or just lose your place easily in dense text.
Easy to Install
Go to chrome://extensions, flip on Developer mode, click "Load unpacked." Done in under two minutes.
Open Source
The code is on GitHub. Read it, fork it, or open a PR. Nothing is hidden.
Getting started
- Clone or download the Voice Reader repo from GitHub
- Open Chrome and go to
chrome://extensions - Toggle Developer Mode on (top-right corner)
- Click "Load unpacked" and point it at the Voice Reader folder
- Select any text on a page and click the Voice Reader icon — or right-click and choose Voice Read
First run downloads the voice model (~63 MB, one-time). After that it works offline entirely.
Want screenshots of every step? See the full install guide.
Features
- Select & Listen: Highlight any text and hear it read aloud immediately
- Word Highlighting: Each word lights up as it's spoken — useful for following along in long docs
- Focus Mode: Strips out the surrounding page so you can concentrate on what's being read
- Speed Control: Slow it down for a foreign language, speed it up for content you're already familiar with
- Multiple Voices: Pick the voice that's easiest on your ears
- Offline First: No internet required after the initial model download
Who uses it
- Getting through long articles or documentation without eye strain
- Studying — hearing something while reading it helps retention for a lot of people
- Dyslexia and other reading difficulties (word highlighting especially)
- Language learning: hear correct pronunciation while reading
- Listening while doing something else — cooking, commuting with a laptop, whatever
How the offline part actually works
Voice Reader uses Piper VITS, a neural text-to-speech engine originally built for low-resource hardware. Combined with WebAssembly, the model runs inside the browser itself — no external API calls, no round trips to any server. Your text goes in, audio comes out, all locally.
The technical details are in the Piper + WebAssembly explainer.
Privacy
Voice Reader has no telemetry, no analytics, and no network requests tied to your usage. It can't log what you read because it never sees it — everything stays in the browser tab.
- Runs 100% offline after model download
- No data sent to any server, ever
- No account or login required
- Nothing stored in the cloud
Why not just use Speechify?
Speechify's free tier caps how much you can listen to per month, and the paid plan runs $139/year. Voice Reader is free, open source, and doesn't gate any features. The audio quality is comparable — both use neural voices. The difference is one sends your text to their servers; the other doesn't.
Side-by-side comparison with Speechify and other paid alternatives.