Free text to speech reader with natural voices

What a text reader is actually good for

Most people meet text to speech through accessibility software and assume that is the whole story. It is not. Hearing words read back is useful to anyone whose eyes are tired, whose attention has drifted, or who has read the same paragraph four times without taking any of it in.

The technology is simple enough to describe. Your device already holds a set of voices, installed by the operating system. The browser hands your text to one of them and it speaks. Nothing is generated from scratch, nothing is downloaded, and on the voices that run locally, nothing leaves your computer.

Proofreading, which is the use people underrate

If you write anything at all, this is the reason to bookmark the page.

Your eye does not read what is on the page. It reads what it expects to be there, especially in your own writing, which you have already read eleven times. That is why the missing word survives four rounds of checking and then leaps out the moment somebody else opens the document.

Your ear has no such loyalty. Listening to your own writing exposes:

• Missing and duplicated words. The the that your eye smooths over is unmissable out loud.
• Sentences that run out of air. If the voice cannot get through it in one breath, neither can your reader.
• Repeated openings. Four paragraphs beginning with However become obvious immediately.
• Clumsy rhythm. A sentence that reads fine and sounds wrong is usually a sentence that needs rewriting.
• Meaning you assumed. Hearing it read plainly reveals the sentence that only made sense inside your head.

Slow the speed slightly for this. Around nought point nine gives you time to react before the voice has moved on, and following the highlighted words while listening catches far more than either alone.

Studying and getting through reading

Listening while reading engages two routes into the same material, and for a lot of people that makes it stick better than reading alone. It also stops the drift where your eyes travel down the page while your mind goes somewhere else entirely, because the voice keeps moving and pulls your attention with it.

For dense material, slow it down and read along. For reading you simply need to get through, push the speed to one point three or one point five. Most people adapt to faster speech within a couple of minutes and stop noticing it. Anything past about one point eight starts costing comprehension for most listeners, so speed for its own sake is not the goal.

Accessibility, and everyone else

For readers with dyslexia, hearing text while seeing it highlighted removes a great deal of the decoding effort and leaves more attention for the meaning. For people with low vision or eye strain, it removes the screen entirely. For anyone recovering from a headache or a long day of screens, it is simply a relief.

There is also a quiet everyday use that has nothing to do with need. Listening to a long article while cooking, walking or folding laundry turns dead time into something. Podcast apps built an industry on that idea. Any article can work the same way.

Getting a natural result

Choosing the voice matters most

The voices on your device vary enormously in quality, and the difference between the best and worst one you have is larger than anything the sliders can do. Try several before settling. What sounds pleasant for two minutes may grate over half an hour, so test with a longer passage than you think you need.

The list here comes from what is actually installed on the device you are using, which is why it looks different on a phone and a laptop. If it seems thin, your operating system settings will usually let you install more voices, and they appear here afterwards.

Speed and pitch

Speed is the useful control. Pitch is not, beyond very small adjustments. Push pitch far in either direction and the voice stops sounding human and starts sounding like a novelty, which is fine for a laugh and useless for listening to anything of length.

Punctuation is what makes it breathe

The voice pauses where your punctuation tells it to. Text with commas and full stops in sensible places sounds like speech. Text without them arrives as an unbroken stream, which is exhausting after thirty seconds. If you are pasting something scraped from elsewhere, adding paragraph breaks before you press play makes a bigger difference than any setting.

Why long text breaks other readers, and not this one

Paste a full article into most browser based readers and it stops partway through with no message. People assume the tool is broken. It usually is not. Two browser quirks are behind it, and both are well known to anyone who has built one of these.

First, Chrome truncates a single long utterance rather than speaking all of it. Hand it four thousand characters and you get some of them. Second, Chrome stops speaking after roughly fifteen seconds of continuous output unless something nudges the queue along.

Tools that pass your whole text over in one go hit both problems. Here the text is split into sentences and queued in order, and the queue is kept awake while it runs. That is why a long piece reaches its end, and it is why sentence structure in your text quietly matters to how smoothly it plays.

Saving the audio, and the honest caveat

The speech API in browsers has no method for handing you the sound as a file. It speaks, and that is all it offers. Any site claiming a simple download from browser voices is doing what this page does, or generating the audio somewhere else entirely.

The approach here is to record the tab while it speaks. That works in Chrome and Edge on a computer. When the sharing box appears you must choose This tab and switch on Share tab audio, or you will get a file with nothing in it. The result is WebM audio, which plays in browsers and most modern players, and can be converted to MP3 if you need one.

It is a workaround rather than an elegant feature, and calling it that is more useful than pretending otherwise.

Where the text goes, and where it does not

Most voices installed on your device do their work on your device. Your text goes to the voice, the voice speaks, and nothing travels. That matters if what you are listening to is a draft contract, a private letter, a medical note or something you have not published yet.

A handful of voices are different. The ones named after the browser maker often run in the cloud, which is frequently why they sound better. Those need a connection, and using them means your text is sent away to be spoken. Every voice in the list here is labelled so the choice is yours to make knowingly. Plenty of readers offer both kinds and mention neither.

Nothing is uploaded by this page itself. There is no account, no character limit sold as a premium tier, and no watermark spoken over your audio.

A few things worth knowing before you rely on it

• Abbreviations are guessed at. Dr, St and Mr are usually handled. Industry acronyms often are not, and get spelled out or mangled.
• Numbers can surprise you. Dates, prices and version numbers are read according to rules that do not always match yours.
• Names are unreliable. Unusual spellings and names from other languages are frequently pronounced badly, and there is no fix beyond respelling them phonetically in the text you paste.
• The voice cannot infer emphasis. Sarcasm, irony and rhetorical questions all arrive flat, which is occasionally revealing about your writing.
• Mobile support is thinner. Phones typically have fewer voices and stricter rules about audio playing in the background.

If the words you want read aloud are currently locked inside a screenshot or a scanned page, our image to text tool will pull them out first. For the opposite journey, turning a recording or your own voice into text, use speech to text, and the rest of the text and speech tools cover what usually comes next.