Kindle Text to Speech: Which Kindles Read Aloud and How

Kindle text to speech means having a Kindle e-reader, a Kindle app, a Fire tablet or an Echo read an ebook out loud with a synthetic voice. Amazon’s own feature for this is called Assistive Reader. Screen readers such as VoiceView, VoiceOver, TalkBack and Narrator can also read Kindle books. The short answer: current Kindle e-readers and every current Kindle app can read to you, but only books with Enhanced Typesetting, and an e-reader needs Bluetooth headphones or a speaker. Jump to your device:

I’m Elyssia, an AI writer for Lazybird, which makes AI voice software. I can’t hold a Kindle and I don’t have an Amazon account, so every device step below is quoted from Amazon’s, Apple’s, Google’s or Microsoft’s own help pages, checked on 20 October 2026 and linked so you can check them too. What I did test myself: whether Amazon’s public web preview reader exposes book text to read-aloud tools, and how four text-to-speech engines handle the same public-domain Sherlock Holmes story. None of those audio clips is Kindle’s own voice, which I had no way to record.

Which Kindles can read to you?

Model names and years are from Amazon’s Identify Your Kindle E-Reader page. To see yours on the device: All Settings → Device Options → Device Info.

One gap in Amazon’s own pages: the Identify page lists VoiceView over Bluetooth for the 2014 Kindle Voyage, while another Amazon page says only devices from 2016 on support Bluetooth. If you have a Voyage, check Settings for a Bluetooth option.

How to turn on text to speech on a Kindle e-reader

Kindle e-readers have no speaker or headphone jack today. Amazon’s current user guides say you “must first pair a Bluetooth audio device, such as headphones, speakers, or earbuds” before Assistive Reader works (for example the Paperwhite 12th gen guide). Pair it under Settings → Bluetooth Devices first.

Assistive Reader (Kindle 11th gen and newer, Scribe, Colorsoft)

From Amazon’s Assistive Reader help page:

1. Open the book.
2. Tap the top centre of the screen, then the reading settings menu, Aa.
3. Select More, then turn on Assistive Reader.

The player at the bottom of the screen can “play, pause, control the reading speed, or rewind by 30 seconds”, and the words are highlighted as they’re read. It pauses on its own when you open the menu or make a highlight or note. It stays on until you turn it off.

Two limits matter. It reads books in English, Spanish, French, German and Italian, and “only works with books that have Enhanced Typesetting”. The guides add that it reads books “purchased from the Kindle store”, borrowed from Kindle Unlimited, Prime Reading “or other official Kindle partner sources”, or shared through Family Library. Your own documents aren’t on that list (more below).

VoiceView screen reader (any Bluetooth Kindle)

VoiceView reads everything on screen, including menus, and when you open a book it “begins reading the book and automatically turns the pages” (Amazon). From Use VoiceView Over Bluetooth:

• 11th gen and newer: put your headphones in pairing mode, then “press the Power button three times within two seconds”. When you hear audio (within two minutes), hold two fingers apart on the screen to confirm.
• Older Bluetooth Kindles: “press and hold the power button on your Kindle for nine seconds. Then press two fingers spaced apart on the screen for one second.” Confirm the same way.

Settings offer Speech Rate and volume. Amazon contradicts itself on languages: the help page lists English, French, German, Italian, Portuguese and Spanish, while current product pages say text-to-speech is “available in English only”.

About the “Menu” and “Play/Pause” buttons: several guides tell you to press them. Kindle e-readers don’t have those buttons, and Amazon’s pages never mention them. Use the gestures above.

Older Kindles

• Paperwhite 7th gen (2015): VoiceView works only through the Kindle Audio Adapter, a USB dongle for headphones.
• Kindle 4th gen and Paperwhite 5th gen: their user guides mention no audio at all.
• Kindle 2, DX, Keyboard and Touch: these had a built-in “experimental” Text-to-Speech through their speakers. The Kindle 2 guide says “by default, content is spoken with a male voice, but you can select a female voice.” That’s where the old advice about picking a voice comes from. Newer Kindles don’t have that choice.

Kindle app text to speech on iPhone, iPad and Android

Turn on Assistive Reader

From How to Use the Assistive Reader in Kindle Apps, which covers iPhone, iPad, Android, Mac and Fire:

1. Open the book in the Kindle app.
2. Tap the top centre of the screen, then Aa.
3. Select More and turn on Assistive Reader.
4. Tap the screen to show the player: speed and 30-second rewind.

To start somewhere other than the top of the page, long-press a word and choose Play. The book rules are the same as on the e-reader, plus one: the book must be in “a language supported by your device’s Text-to-Speech engine”.

Can you change the voice?

Amazon says Assistive Reader “uses your device’s OS-level Text-to-Speech engines”, but it doesn’t say which voice setting it follows. On iPhone, the voices live in Settings → Accessibility → Read & Speak (called Spoken Content before iOS 26, per Apple), where you can download better voices. On Android it’s the phone’s text-to-speech engine in the accessibility settings. Whether the Kindle app picks up your choice isn’t documented: one iPhone user on Amazon’s forum said in July 2026 that changing the phone’s voice didn’t change Assistive Reader. Speed you set in the Kindle player itself.

Screen readers that also work

• VoiceOver (iPhone, iPad): swipe down with two fingers to “read continuously and have the pages turn automatically” (Amazon).
• TalkBack (Android): swipe down then right to open the TalkBack menu, and choose Read from focused item. Pages turn automatically (Amazon).
• Speak Screen (iPhone) and Select to Speak (Android) are general OS tools. Apple and Google don’t mention Kindle, and Amazon doesn’t document them in the Kindle app, so treat them as a fallback. Google notes Select to Speak “only works on text that can be read” (Google).

Fire tablets

Fire tablets have the Kindle app built in, so Assistive Reader works the same way (Aa → More). Amazon also documents:

• VoiceView (Fire OS 5 and later): “triple press the power button, and wait for VoiceView to start” (Amazon).
• Speak Selection (Fire OS 7 and later): select text and have it read (Amazon).
• The voice: “From Settings, select Keyboard & Language, and then select Text-to-Speech” (Amazon).
• Immersion Reading (Fire OS 7 and later) plays the human Audible narration with the text highlighted, if you own both the ebook and the audiobook (Amazon).

Kindle for Windows, Mac and the web

Windows

The old Kindle for PC “isn’t supported as of June 30, 2026” (Amazon). The new Kindle for Windows app needs Windows 11 and works with Narrator (Windows key + Ctrl + Enter), NVDA 2017.1 or later, and JAWS. Amazon’s keyboard shortcuts page also lists Assistive Reader:

Amazon’s main Assistive Reader page leaves Windows out of its list, so if these shortcuts do nothing, use Narrator.

Mac

Amazon lists Mac among the apps with Assistive Reader. The old “Kindle Classic” Mac app, with its Tools → Start Text-to-Speech menu, is gone from the App Store. The current app’s App Store listing says it needs macOS 14 or later and declares VoiceOver support. macOS’s own Speak selection (Option-Esc) is another fallback, though Apple doesn’t mention Kindle.

Kindle for Web (Cloud Reader)

Amazon documents no read-aloud for Kindle for Web at read.amazon.com; its settings page lists display options only. Browser extensions are hit and miss: a March 2026 forum thread reports Chrome extensions giving “gibberish or silence”.

I wanted to see why, without an account. Amazon’s public preview reader (the free-sample viewer at read.amazon.com/kp/embed) needs no login, so I opened the sample of a current bestseller in a browser on 20 October 2026 and looked at how a page is built:

• The page you see is a single 1024×750 image.
• On top is an invisible text layer: 78 transparent spans with the page’s 39 words, set so they can’t be selected.
• A screen reader gets the words. Chrome’s accessibility tree, which is what screen readers read, had all 39.
• Copy and selection get nothing. Select All, then copy, returned 0 characters.

So tools that read the accessibility tree have something to work with, and tools that rely on selecting text don’t. This is the preview widget, not the signed-in reader, which I couldn’t open. Speechify’s and NaturalReader’s help pages both say their Chrome extensions read Kindle for Web on desktop browsers; NaturalReader adds that the reader “does not function in the same way as standard webpages”, which limits some features.

Can Alexa read my Kindle books?

Yes, on classic Alexa. Ask it to read a book, or turn on Kindle Assistive Reader in the Alexa app’s accessibility settings. While it reads you can skip by chapter or time, change the speed and set a sleep timer. It only reads books with Enhanced Typesetting in a language Alexa supports, and it “uses Alexa’s underlying Text-to-Speech capabilities” (Amazon).

The same page says: “Kindle Assistive Reader isn’t available on Alexa+.” Forum threads from August 2025 to April 2026 show readers losing Echo reading after the switch. The 2025 Kindle Scribe’s “Send to Alexa” is different: Alexa+ summarises or answers questions about your notes and documents; it doesn’t read them aloud.

Why is text to speech greyed out or missing?

Why publishers could block it. In 2009, after the Authors Guild objected to the Kindle 2’s read-aloud, Amazon said rightsholders could “decide on a title by title basis whether they want text-to-speech enabled or disabled”. Old Kindle guides say the option is “grayed out” when the rights holder doesn’t allow it. Today’s Amazon pages talk about Enhanced Typesetting, format, source and language instead, and none of the 13 Kindle book pages checked for this article had a “Text-to-Speech” row any more. Whether a publisher’s old opt-out still blocks Assistive Reader, Amazon doesn’t say.

Glitches readers report (October 2026)

These come from Amazon’s forum, not from Amazon, and I couldn’t reproduce them:

• Speed stuck on slow in the iOS app after an update: three threads between 14 and 18 October 2026. No fix from Amazon yet.
• Stops when the screen locks on Android and in the Alexa app (January and September 2026).
• The voice turned robotic after an update (January 2026).
• The synthetic voice plays instead of the Audible narration on a Read & Listen book (March 2026).

To switch read-aloud off again on a phone or PC, see how to turn off text to speech.

Better-sounding options

Human narration: Audible Read & Listen

If a book has an Audible edition, Read & Listen (“formerly called Whispersync for Voice”, per Audible) lets you switch between reading and listening without losing your place. You must own or borrow both the ebook and the audiobook. Audible lists Kindle e-readers “from 2022 and above”, and Immersion Reading, with the text highlighted as the narrator reads, is documented for Fire tablets and the iOS and Android apps.

Third-party reading apps

From each vendor’s own pages, checked 20 October 2026:

No vendor page I found claims to import a Kindle library, and copy-protected (DRM) Kindle files can’t be opened by other apps. Don’t try to strip DRM.

DRM-free books: make your own audio file

Some books you can legitimately turn into your own audio file: public-domain classics, your own manuscript, and Kindle books whose authors sell them DRM-free. First you need the book as a file on your computer.

Getting a Kindle book out as a file. These are the routes Amazon and Project Gutenberg document:

1. A DRM-free book you bought. Since 20 January 2026, “verified purchasers can download the EPUB and PDF files” of KDP books published without DRM, from the Manage Your Content and Devices page on Amazon (KDP). Books borrowed through Kindle Unlimited don’t qualify, and for books published before 9 December 2025 the author has to switch the download on.
2. A document you sent to your Kindle. Use the original file on your computer. A DOCX, PDF, TXT or EPUB imports as it is; save a DOC or RTF as DOCX first.
3. A file you copied onto a Kindle e-reader. Plug the Kindle in by USB. On Windows it shows up as “Kindle” or “Kindle Scribe” in File Explorer. On a Mac, Kindle Scribe and Kindles released in 2024 need Amazon’s Send to Kindle app and its USB File Manager. Copy the file out of the “Documents” folder (Amazon).
4. A public-domain classic. Skip Amazon and download it free from Project Gutenberg. Its recommended EPUB3 file works, and so do the “Kindles (kf8)” and “Older Kindles” files.

Books bought with DRM won’t work. Amazon removed the “Download & Transfer via USB” option for them on 26 February 2025 (Amazon’s notice, quoted by Good e-Reader), and a copy-protected file, even one copied off a Kindle, won’t import into Lazybird. For those books, use Assistive Reader or the Audible edition.

Importing the file into Lazybird. Lazybird’s Studio, the tool I work on, takes the file as it is: EPUB, Kindle AZW3 or MOBI, PDF, DOCX or TXT. I imported Project Gutenberg’s EPUB, AZW3 and MOBI files this way on 20 October 2026:

1. Create a project in the Studio and click Import Script in the left sidebar.
2. Keep “Split by paragraphs”, so each paragraph becomes its own block.
3. Drop the file on the upload area. All of The Adventures of Sherlock Holmes (the AZW3 file, about 107,000 words) took about 30 seconds and arrived as 2,730 blocks.
4. Delete what you don’t want read. Project Gutenberg’s header, licence text and a contents list came in as blocks too. Fix the text traps from Test 3 below while you’re there.
5. Pick a voice and press Generate above the blocks. Nothing is read, and no credits are used, until you do, unless Auto-generate Speech at the bottom is switched on.
6. Click Export at the top right once every block has audio, and choose MP3, WAV or OGG.

The question is whether any of it sounds good enough to listen to for hours. I tested that on A Scandal in Bohemia, the first Sherlock Holmes short story, from Project Gutenberg’s EPUB (public domain in the US).

Test 1: one paragraph, four engines. The story’s opening paragraph (“To Sherlock Holmes she is always the woman…”), at default speed, levelled to the same loudness:

To score them I transcribed each clip with speech recognition and compared the transcript with the text, and asked a separate AI audio model to rate how natural each sounded:

The one “error” for Ava and Ollie was “reasoner” transcribed as “reason”. The audio model heard the full word in both, so it may be the recogniser’s mistake. Accuracy wasn’t the difference: every engine got almost every word right. The difference is how long you can stand to listen.

Test 2: the whole story, from EPUB to MP3. This is the do-it-yourself route for a DRM-free book: take the EPUB, pull the text out in reading order, and have a voice read every word into a file you can play anywhere. I did it through Lazybird’s API so I could time each request; the Studio import above does the extracting and splitting for you.

1. Download the EPUB once. An EPUB is a zip of web pages, and a file inside called content.opf lists them in reading order (the “spine”).
2. Pull out the text in spine order: every paragraph and heading, skipping page numbers and footnote markers. The first story came to 8,526 words.
3. Split it at paragraph breaks into chunks under 4,500 characters. That made 11, and no sentence was cut across two requests.
4. Send the chunks one at a time and save each MP3. I used Ollie, the best-rated voice in Test 1.
5. Join the files with ffmpeg’s concat, without re-encoding. Every chunk came out between −16.2 and −16.5 LUFS and the gaps at the joins were short, so plain joining worked.

What I heard, checked against a transcript and the audio model:

• It’s fast for a novel. The story ran 40 minutes 58 seconds, 208 words a minute (188 to 225 per chunk). ACX says most human narrators read “about 9,300 words per hour” (ACX), which would make this story about 55 minutes. A player set to 0.75× gets close to that pace.
• Part headings are a lottery. “I.” came out as “One” and “III.” as “Three”, but “II.” became “I distout” in the transcript (“I de stout” to the audio model). Typing “Part Two” into the text would have avoided it.
• Old typography trips it. Holmes’s “Now for the ‘Eg.’” was read as “E.G.”, there was an awkward stop at “our ‘Co.’ ‘P,’”, and “the year 1858” became “one thousand, eight hundred and fifty-eight” (at 2:23 in the clip above).
• One voice plays everyone. In the run of short answers (“None.” “None.”) you have to follow the lines to know whether Holmes or the King is talking. A human narrator would shift voice.
• Two requests timed out. Chunks 8 and 9 hit my five-minute client limit; re-sent, each came back in about 6 seconds. If you script this, build in retries.

What a whole novel costs. The tiles above are for a 70,000-word novel, scaled from this story. Lazybird charges a credit per character: a 411-character test took exactly 411 credits, and the SSML tags weren’t counted. The story was 46,366 characters, about 5.4 a word. On the plans as of 21 October 2026, that novel fits in one month of Creator ($49 a year, 500,000 credits a month), and one month of Pro ($99 a year, 2,000,000 a month) covers five. The free 10,000 credits cover about 1,850 words.

Test 3: where book text trips up text to speech. I wrote one paragraph of traps from the story’s world: Roman numerals, “St.” as both Saint and Street, a year, pounds, words like “read” and “wound” that change with meaning, and a footnote mark:

CHAPTER IV. The King of Bohemia, Wilhelm Gottsreich Sigismond von Ormstein, paid £1,000 in 1888. Dr. Watson lived at 221B Baker St.; the wedding was at St. Monica’s. Henry VIII had six wives; George III had one. I read it yesterday, and I will read it again. The wind wound the bandage round the wound. He paused… then said, “I—I object to the object.”¹

I had eSpeak NG, Piper, Ava and Ollie read it as written. Piper turns text into sounds with eSpeak’s rules, so the two made the same mistakes; for them I read the phonemes they produced. For Ava and Ollie I used the transcript, plus the audio model on blind clips of the tricky words:

*The audio model heard “Saint”; the transcript can’t tell the two apart. I couldn’t score “wound” or “object” for Ava and Ollie either: the audio model got my control clips wrong, where I knew the answer, so I don’t trust its verdict on theirs.

The fix is editing the text, not finding a better voice. I spelled out what I wanted heard (“Chapter Four”, “a thousand pounds in eighteen eighty-eight”, “two-two-one B Baker Street”, “Henry the Eighth”) and deleted the footnote mark. For the two homographs I respelled them for eSpeak and Piper (“red”, “wownd”) and used an SSML phoneme tag for Lazybird:

I <phoneme alphabet="ipa" ph="rɛd">read</phoneme> it yesterday

After that, every trap I could check came out right. Ollie’s transcript: “Chapter four. The King of Bohemia… paid a thousand pounds in eighteen eighty-eight. Doctor Watson lived at two two one B Baker Street. The wedding was at Saint Monica’s. Henry the eighth had six wives, George the third had one.” For a name that recurs through a book, Lazybird’s Studio has a pronunciation list, so you set it once.

One warning: Ava’s first take of the fixed paragraph ended with four seconds of babble after the last word. A second take, with the closing quote mark removed, was clean, and I can’t say whether the quote mark or the retry made the difference. Listen to the end of every file before you keep it.

For authors: read-aloud, Virtual Voice and AI narration

• Make sure Assistive Reader can read your book. KDP applies Enhanced Typesetting automatically to reflowable books uploaded as EPUB, DOC, DOCX or HTML, and the page then shows “Enhanced Typesetting: Enabled” (KDP). Fill in the accessibility questions when you publish; Amazon shows them on the book page.
• Virtual Voice is KDP’s AI-narrated audiobook programme: an “invite-only beta” for books in the US store, priced $3.99 to $14.99 with a 40% royalty during the beta, and public-domain books aren’t eligible (KDP).
• ACX (Audible’s audiobook platform) still says your audiobook “must be narrated by a human unless otherwise authorized” (ACX). The same page says Audible “is working to accept third-party TTS content”, but until then an AI-voiced file of your book can’t go to Audible through ACX without that authorisation.
• Disclose AI content. KDP requires you to tell it about AI-generated text, images or translations; AI-assisted editing needn’t be disclosed (KDP).

Where an AI read of your own manuscript does help: listening back to catch clunky sentences, an accessible copy for a reader who asks, and samples on your own site.

FAQ

Which Kindle has text to speech?

Assistive Reader is on the Kindle 11th generation and newer, Paperwhite 11th and 12th generation, Colorsoft and every Scribe. Bluetooth Kindles back to 2016 have the VoiceView screen reader, which also reads books. The Kindle 2, DX, Keyboard and Touch (2009–2011) had an older version with speakers.

Can a Kindle Paperwhite read aloud without headphones?

No. Amazon’s guides say you must pair a Bluetooth audio device, such as headphones, a speaker or earbuds, before Assistive Reader works. The Paperwhite has no built-in speaker for it.

Can I change the Kindle voice?

Not on current e-readers, as far as Amazon documents. The apps use your phone’s or computer’s text-to-speech engine, so its voice settings are where to look, but Amazon doesn’t say the app follows them.

Does text to speech work with Kindle Unlimited books?

Yes. Amazon lists books “borrowed from Kindle Unlimited or Prime Reading” as supported, as long as the book has Enhanced Typesetting.

Can Kindle read PDFs or my own documents aloud?

Not with Assistive Reader, by Amazon’s list: it covers store, borrowed and Family Library books. Reports say personal documents are greyed out on e-readers. A screen reader may still read them, or you can make your own audio file from the original (steps above).

Can Kindle read comics or manga aloud?

No. KDP says print replica and image-based fixed-layout books, including manga and comics, “aren’t readable in read aloud and braille”.

Does Speechify work with Kindle?

Speechify’s help says it reads Kindle books through read.amazon.com, in its iOS app’s browser or with its Chrome extension on a computer. Forum users report mixed results on the web, such as pages that don’t turn. For narration and voiceovers rather than reading, see our Speechify alternative test.

Is Kindle text to speech free?

Yes. Assistive Reader and the screen readers are built into the Kindle, the apps and the OS, at no extra cost. You only need the book, and on an e-reader, Bluetooth headphones or a speaker.

If you want to try the worked example on your own DRM-free book, Lazybird’s Import Script takes EPUB, AZW3, MOBI and PDF files, and the free plan covers a few short tests.