What you’ll be able to do

You can get almost any bit of text, a PDF, an article, an ebook, a Word doc, read out to you in a voice that doesn’t make you want to switch it off after two minutes. That’s the whole point of this guide.

Once you’re set up, you can listen to a report whilst you’re doing the washing up, get through a long article on the train, or have a book read to you when your eyes are too tired to keep reading it yourself. Some of this uses proper AI voices, the kind that sound close to human. Some of it just uses the read aloud tools already sitting on your phone or laptop, which honestly are better than most people realise.

Before you start

Here’s the bit that surprises people: you probably don’t need to install anything new. If you’ve got a phone, a modern browser, or a copy of Word, you already have a read aloud tool somewhere on your device. Most of this guide is about finding it and tuning it, not buying something.

Now the less fun bit, rights. If it’s your own document, your own notes, an article you’re allowed to read, that’s fine, read away. If it’s a book you’ve bought, reading it aloud to yourself with a text to speech tool is fine too. What’s not fine is taking someone else’s book, running it through an AI voice, and sharing that recording with other people or posting it online. That’s not your content to redistribute, however nice the narration sounds. Keep it personal and you’re on safe ground.

Get set up

Start with what’s already on your device, because it’s free and it’s genuinely good now.

On Windows, Microsoft Edge has a Read aloud feature built right in. Open any web page or PDF in Edge, and there’s a button (usually in the address bar or right click menu) that starts reading it to you with a natural sounding voice. Works on articles, works on PDFs, no faff.

If you’re on a Mac or iPhone or iPad, look for Spoken Content, sometimes shown as Speak Screen, in your accessibility settings. Turn it on once and you can get any screen read aloud with a swipe gesture. It’s quietly one of the best accessibility features Apple ships, and most people never turn it on.

Android has something similar called Select to Speak, plus a Reading mode in some apps and browsers that strips out clutter and just reads the text.

Using Word a lot? There’s a Read Aloud button built in, tucked under the Review tab. Click it and it reads your document back to you, which is also a sneaky good way to catch typos, your ear picks up mistakes your eye skips over.

If you’re into audiobooks and ebooks specifically: Kindle books can be read aloud through Alexa if you’ve got an Echo or a Fire device. Google Play Books has read aloud for a lot of its titles too. Worth checking before you go looking elsewhere.

Once you’ve tried the free built-ins and want something that sounds a bit more polished, that’s when it’s worth looking at paid AI voice apps like the ElevenLabs Reader app, Speechify, or NaturalReader. They tend to offer more natural sounding voices and more control over pacing, but they cost money, so try free first and only pay if you’re actually going to use it a lot.

Try it yourself

If you want an AI assistant like ChatGPT, Claude, or Gemini to help prepare text before you read it aloud (cleaning it up, breaking it into chunks, adding pauses) here are prompts that do different jobs.

To get a long article turned into something easier to listen to in one sitting:

Take the following article and rewrite it slightly so it reads well out loud. Keep the meaning exactly the same, but break up any long or complicated sentences, and add natural pauses where a narrator would take a breath. Don't add new information. Here's the text:

[paste your article]

To get a document split into sensible chunks for listening in short bursts, say on a commute:

Split this document into sections of roughly 3 to 5 minutes of spoken reading time each. Give each section a short title so I know where I am. Don't summarise or shorten the content, just divide it. Here's the text:

[paste your document]

To turn dense or technical writing into something that sounds natural when read aloud, without dumbing it down:

Rewrite this passage so it sounds natural when spoken aloud by a text-to-speech voice. Keep all the facts and the level of detail the same, just adjust sentence structure and wording so it flows when heard rather than read. Here's the text:

[paste your passage]

Once you’ve got the text prepped, hand it over to whichever tool is actually going to speak it, Edge’s Read aloud, Spoken Content, Word, or one of the paid apps. The AI assistant is doing the writing prep, not the speaking, in most of these setups.

Check the result

Don’t just assume it’ll be fine, actually test it before committing to a long document or a whole book.

Pick a short paragraph, maybe two minutes worth, and listen to it at your normal working volume, in the environment you’ll actually use it in. Kitchen with the extractor fan on. Train with background noise. That sort of thing.

Ask yourself: could I listen to this voice for an hour without getting irritated? That’s the real test, not whether it sounds impressive for thirty seconds.

Check the pacing. Most tools let you adjust speed, and for long documents, 1.2 to 1.5 times normal speed tends to be the sweet spot, fast enough to save time, slow enough that you’re not straining to keep up.

Listen for how it handles names, numbers, and abbreviations, these are where text to speech tools often trip up. If it mangles a name every time it appears, that’s going to grate over a whole document.

And check whether it pauses in sensible places, at full stops, at paragraph breaks, rather than ploughing straight through everything in one flat stream.

If all of that feels comfortable, you’re set. If something’s bugging you, small tweaks usually fix it, which brings us to the next bit.

If it doesn’t work

First, the privacy point, because it matters more here than people expect. If you’re pasting text into an AI assistant to get it rewritten or chunked up, don’t paste anything containing passwords, financial details, medical records, or anything genuinely private, even if it feels like a harmless step. Treat that text box the same way you’d treat an email to a stranger. If the document has sensitive content, use the built-in tools that stay on your own device (Edge, Spoken Content, Word’s Read Aloud) rather than sending it anywhere at all.

If the voice sounds robotic or flat, the fix is usually one of: switching to a different built-in voice (most devices have several to choose from, and they vary a lot), adjusting the speed, or moving up to one of the paid AI voice apps, which generally sound more natural than the free system voices, at the cost of a subscription.

If it’s mispronouncing names or technical terms constantly, that’s a known limitation of most text to speech tools, there’s no perfect fix, but breaking the word up phonetically in the text sometimes helps, and some paid apps let you add custom pronunciations.

If nothing’s reading aloud at all, check the basics first: is the accessibility feature actually turned on, is the correct voice pack downloaded, is your volume actually up. Sounds obvious. It’s the cause about half the time.

And if you’re trying to get a whole book read and it’s refusing or looking odd, check whether it’s a rights issue rather than a technical one, some ebook formats and DRM protected files won’t play nicely with read aloud tools, which is by design, not a bug.

Keep exploring

If you’ve got audio you need turned into text instead of the other way round, there’s a guide on this site called How to get AI to transcribe an audio recording.

If you’re drowning in long documents and want the short version before deciding whether to listen to the whole thing, try How to get AI to summarise a document.

And if you need something translated whilst it’s being spoken, in a meeting or a conversation, there’s How to get AI to translate a live conversation.

Sources and review notes

Review date: 15 September 2026 — every source above was opened and checked against the text on that date.