The libraryLibrary

I already sell

How to turn a PDF into an audiobook with AI: the complete guide

Sonia BéUpdated November 20, 20269 min read

You have a PDF (a guide, a course, an ebook) and you'd like to hear it, or let your customers listen to it. To listen yourself, your computer or phone can already read it aloud. To publish a real audiobook, you need a bit more method: text adapted for the ear, a well-chosen voice, clean files and the rules of the platform you're aiming for. Here are both paths, step by step, with or without Lektify.

Listening to a PDF or publishing an audiobook: what do you need?

It all depends on who will be listening.

  • Just you: you want to listen to a document on a walk or review it with your eyes closed. Built-in read-aloud is enough, with nothing to export.
  • Your readers or customers: you want to offer them an audio version. Then you need audio files, chapter by chapter, or narration built into your ebook, with a voice you've chosen.

Audio also helps people who find reading hard: a meta-analysis published in 2018 concludes that text-to-speech may help students with reading difficulties understand a text (Wood et al., 2018). On reading comfort, see also what research says about fonts for dyslexia.

MethodFor whomWhat you getIts limit
Built-in read-aloudYouInstant listeningNo file, the device's voice
A. Text-to-speech toolYour readers, anywhereOne MP3 per chapterChecking and publishing are up to you
B. Platforms (Spotify, Amazon)Their listenersAn audiobook in their catalogLimited languages, countries and accepted voices
C. Narration in your ebookYour readersText and audio in one placeThe audio plays inside the ebook

How do you listen to a PDF for free?

With the reading features already on your devices:

  1. On a computer, with Microsoft Edge. Open your PDF in Edge, on Windows or Mac, and start Read aloud. The passage being read is highlighted (Microsoft).
  2. With Acrobat Reader, Adobe's free app: View menu, then Read Out Loud. You can have the current page read, or the whole document (Adobe).
  3. On iPhone. Go to Settings, Accessibility, Read & Speak, and turn on Speak Screen. Open your PDF, then swipe down with two fingers from the top of the screen: your iPhone reads all the text shown (Apple).
  4. On Android. In Settings, Accessibility, open Select to Speak and turn on its shortcut. Start it with a two-finger swipe up, then tap Play to hear everything (Google).

Two limits. These features read, but they don't make any file you can share. And they need real text: if you can't select a sentence in your PDF, it's an image (a scan), and it needs text recognition (OCR) first. Google Drive can do this: right-click the file, then Open with, Google Docs. The file should be 2 MB or smaller, and lists, tables, columns and footnotes are not likely to be detected (Google Drive).

How do you prepare your PDF for the ear?

Text written for the eye doesn't always sound good. Before you generate anything, do this review, chapter by chapter:

  1. A clear title for each chapter. By ear, there's no table of contents and no page number. Announce each chapter in its first words, as Spotify's guidelines ask (Spotify for Authors).
  2. Tables as sentences. Read aloud, a table becomes a string of cells with no landmarks. Keep the main idea in one or two sentences and put the full table in an appendix.
  3. Numbers, dates, acronyms and units. Synthetic voices may mispronounce phone numbers, currencies, dates, web addresses or unit abbreviations (ElevenLabs). Write them the way you want them heard: "15 dollars" rather than "$15", "kilometers" rather than "km".
  4. Cross-references and notes. "See page 12" or "below" mean nothing when spoken. Replace them with the section's name, and work useful notes into the text.
  5. Images and captions. If an image carries information, describe it in one sentence. Otherwise, leave it out.
  6. Long sentences. Split them. A sentence that reads fine in silence can leave a voice out of breath.

Keep this revised text separately: you'll use it whatever method you choose.

Which voice should you choose: synthetic, cloned or human?

VoiceStrengthsLimits
Synthetic voice (catalog)No recording, plenty of voices and languagesIt isn't your voice; it must be disclosed on Spotify
Cloned voice (yours, made with AI)Your readers recognize your voiceYou need a clean recording to start, and you should only clone your own voice
Human voice (you or a narrator)The most natural; the only one accepted by default on ACX (Audible)Equipment, time and editing

For a practical guide, a well-chosen synthetic voice often does the job; for a personal story, your own voice builds more connection.

Method A: how do you create one MP3 file per chapter?

Let's take Studio, from ElevenLabs, as an example: its audio projects are designed for long-form narration, podcasts and audiobooks. Other text-to-speech tools follow the same principle.

  1. Import your text. Studio accepts PDF, DOCX, TXT, HTML or a URL, and recommends EPUB; chapters are detected automatically when you import a document (ElevenLabs). Import your revised text rather than the raw PDF if you can.
  2. Choose the voice, then add tricky words to a pronunciation dictionary.
  3. Generate and listen chapter by chapter, fixing things as you go.
  4. Export each chapter on its own, or the whole project as a ZIP file with one file per chapter.
  5. Check the export quality. The Free, Starter and Creator plans export 128 kbps MP3; you need Pro or higher for 192 kbps MP3 or 44.1 kHz WAV. Spotify, for its part, asks for MP3 at 192 kbps or higher, or else WAV or FLAC (Spotify for Authors).

The free plan includes Studio, with 3 projects and 10,000 credits a month, and the pricing page lists a commercial license only from the Starter plan (ElevenLabs). It's great for testing a chapter, less so for publishing.

Method B: can you publish directly on Spotify or Amazon?

Yes, but each route has its own conditions:

RouteSynthetic voiceMain conditions
Spotify for Authors (upload)Accepted, must be disclosedFree, non-exclusive, MP3, WAV or FLAC
Spotify's creation toolsElevenLabs voicesSelf-published authors in the United States, English-language titles, free for up to 5 audiobooks
Virtual Voice (Amazon KDP)Yes, that's the whole ideaInvite-only beta; ebook in English, Spanish, Italian or French; sold in the US marketplace
ACX (Audible)No, unless authorizedResidents of the United States, United Kingdom, Canada and Ireland

Spotify for Authors. Uploading is free, with no limit on the number of titles, and non-exclusive: you're free to distribute elsewhere (Spotify). With a synthetic voice, select “This audiobook uses digital voice narration”, and Spotify tells listeners in the description (Spotify). Its built-in creation tools, which use ElevenLabs voices, are currently limited to English-language titles by authors based in the United States (Spotify for Authors).

Virtual Voice, on Amazon KDP. This option turns an ebook already on sale on KDP into an audiobook narrated by a synthetic voice. The conditions: an invite-only beta, an ebook whose primary language is English, Spanish, Italian or French, and distribution limited to the US marketplace. An ebook that already has an Audible audiobook isn't eligible, and if your ebook is enrolled in KDP Select, its exclusivity terms apply (KDP).

ACX, the route to Audible. ACX is open only to residents of the United States, the United Kingdom, Canada and Ireland (ACX). Audiobooks there must be narrated by a human, unless otherwise authorized. Its technical requirements are a good benchmark even elsewhere: MP3 at 192 kbps or higher, only one chapter per file, and no file longer than 120 minutes (ACX).

Method C: how do you add audio to your ebook with Lektify?

On Lektify, the platform that publishes this blog, your PDF becomes a Lekty'Book: an ebook read online, where text and audio live in the same place.

  1. Import your PDF. In My Lekty'Books, click Import a PDF (50 MB max). Choose Import quickly, which is free, or Import faithfully: AI then rebuilds the pages with tables, boxes, columns or scans (100 free pages a month on Freemium, 300 on Creator and Business, then 1 token per page). Check the detected chapters, confirm that you hold the rights, including for audio narration, and create your Lekty'Book (import guide). Have the Word version? Import that instead: it keeps your headings, tables and links, for free.
  2. Prepare the text for the ear in the editor, using the list above. The narration reads headings, text and callout boxes; it skips quizzes, images and videos.
  3. Choose your narrator. In the editor, open AI Tools, then Audio narration. Under Your narrator, listen to your voices, or click More voices to filter the catalog by gender, pitch, tone, age or accent. Your choice applies to every chapter.
  4. Or create My voice. The Create my voice link has you read a consent sentence aloud, then three short texts: about 2 minutes of recording. Your cloned voice then reads your chapters, speaks five languages (French, English, Spanish, German and Portuguese), and can only be used from your account (My voice).
  5. Generate the audio. The price appears on the Generate audio button: 10 tokens per 1,000 characters read with a catalog voice, 18 with My voice, and 20% more on Freemium (1 token = 1 euro cent). So a 20,000-character chapter costs 200 tokens with a catalog voice on Creator or Business. If generation fails, your tokens are refunded. After an edit, click Regenerate (audio narration, price list).
  6. Your readers listen in the Reader's Audio tab, labeled “AI narration · Chapter”, with 7 speeds from 0.5× to 2× (listening to a chapter), without leaving the Lekty'Book.

Good to know: the account and quick import are free, but narration is paid for in tokens, and the Freemium plan includes none each month (Creator includes 1,000). The audio is made to be played inside the Lekty'Book: for Spotify or Audible, produce your files with method A. And audio is just one way to get your reader involved: interactive ebooks offer others.

How do you check the quality before publishing?

Whatever the method, keep this list handy:

  1. Listen to everything, with the text in front of you. A synthetic voice can stumble over a list or read out a symbol.
  2. Names, acronyms, foreign words. Note the ones that sound wrong, then fix them: a pronunciation dictionary in ElevenLabs Studio, pronunciation rules in Spotify's tools, phonetic edits of a word or short phrase in Virtual Voice (KDP). Otherwise, respell the word the way it's pronounced.
  3. Pace. Listen at normal speed and split any sentence that leaves the voice out of breath.
  4. Consistent volume from one chapter to the next. Audacity, a free, open source audio editor (Audacity), adjusts volume with its Loudness Normalization effect: in LUFS (default -23, the value of the EBU R 128 recommendation) or in RMS (default -20 dB) (Audacity manual, EBU). That -20 dB falls within both ranges: -23 to -18 dB RMS for ACX, which also requires peaks no higher than -3 dB and a noise floor no higher than -60 dB, and -24 to -14 dB RMS for Spotify.
  5. The start and end of each file. Announce the chapter and leave a short silence: Spotify recommends 0.5 to 1 second at the start and 1 to 5 seconds at the end.
  6. A final test with headphones, on a phone, where your listeners will actually be.

Already selling a PDF course? An audio version is one of the additions that help your customers finish. If you then rethink your price, reread our method for pricing an ebook.

So, where should you start?

To listen to your PDF, turn on your device's read-aloud feature. To let others listen, start small: prepare one chapter for the ear, have it read by two or three voices, listen, fix. Before you aim for a platform, check the language, the country and the accepted voices. For your own readers, audio built into your ebook keeps everything in one place.

Want to try it? Create your free account, import your PDF with the free quick import, then generate the narration for a first chapter, with a catalog voice or your own.

Test yourself

Three questions to check what you remember.

Question 1 of 3What does read-aloud in Edge or Acrobat Reader do with your PDF?

Frequently asked questions

How do you convert a PDF into an audiobook for free?

To listen yourself, use the read-aloud features already on your devices: Read aloud in Microsoft Edge, Read Out Loud in Acrobat Reader, Speak Screen on iPhone or Select to Speak on Android. They read the text but don't create a file. To get MP3 files, some AI tools have a limited free plan, such as ElevenLabs Studio (3 projects, MP3 export at 128 kbps).

What is the best AI voice for audiobooks?

There's no single best voice for everyone: have the same page read by two or three voices and keep the one that sounds right in your language. For free options, ElevenLabs Studio has a limited free plan, and the Spotify for Authors creation tools are free for up to 5 audiobooks, currently for English-language titles by self-published authors in the United States. Before you publish, check that your plan allows commercial use.

How do you make an audiobook on Spotify?

With Spotify for Authors, uploading is free and non-exclusive: you upload one MP3, WAV or FLAC file per chapter and a square cover. If you used a synthetic voice, select “This audiobook uses digital voice narration”, and Spotify adds a short sentence to the description to let listeners know.

How do you make an audiobook on Amazon?

There are two routes. Virtual Voice, in KDP, turns an ebook already on sale on KDP into an audiobook narrated by a synthetic voice: it's an invite-only beta for ebooks in English, Spanish, Italian or French, distributed only in the US marketplace. ACX, the route to Audible, requires human narration unless otherwise authorized, and it's open only to residents of the United States, the United Kingdom, Canada and Ireland.

How do you turn a PDF into an audiobook on iPhone?

To listen, go to Settings, Accessibility, Read & Speak, and turn on Speak Screen. Open your PDF, then swipe down with two fingers from the top of the screen: your iPhone reads all the text shown. It doesn't create an audio file, so to share an audiobook you'll need a tool that exports MP3 files.

Sources

23
  1. Microsoft Learn, “PDF reader in Microsoft Edge” (Read aloud feature, Windows and macOS) (accessed November 10, 2026)
  2. Adobe, “How to have your PDF files read aloud to you” (Acrobat Reader, View menu, Read Out Loud) (accessed November 10, 2026)
  3. Apple, iPhone User Guide (iOS 27), “Hear iPhone speak the screen, selected text, and typing feedback” (accessed November 10, 2026)
  4. Android Accessibility Help, “Use Select to Speak” (accessed November 10, 2026)
  5. Google Drive Help, “Convert PDF and photo files to text” (text recognition, files of 2 MB or smaller) (accessed November 10, 2026)
  6. Wood, Moxley, Tighe and Wagner (2018), “Does Use of Text-to-Speech and Related Read-Aloud Tools Improve Reading Comprehension for Students With Reading Disabilities? A Meta-Analysis”, Journal of Learning Disabilities (accessed November 10, 2026)
  7. ElevenLabs, “Best practices” documentation, “Text normalization” section (numbers, dates, URLs and units that get misread) (accessed November 10, 2026)
  8. ElevenLabs, “ElevenCreative Studio” documentation (import formats, chapters, pronunciation dictionaries, per-chapter export and quality by plan) (accessed November 10, 2026)
  9. ElevenLabs, Pricing page (Free plan: 10,000 credits a month, 3 Studio projects; commercial license listed from the Starter plan) (accessed November 10, 2026)
  10. Spotify for Authors, “Uploading audiobooks to Spotify for Authors” (free, non-exclusive, MP3, WAV or FLAC) (accessed November 10, 2026)
  11. Spotify for Authors, “Digital voice narration” (checkbox and disclosure sentence added to the description) (accessed November 10, 2026)
  12. Spotify for Authors, “Audiobook Creation Tools” (self-published authors in the United States, English-language titles, free for up to 5 audiobooks) (accessed November 10, 2026)
  13. Spotify for Authors, “Metadata & Assets Guide” (last updated November 2024; audio specs, one file per chapter, spoken chapter announcement) (accessed November 10, 2026)
  14. Amazon KDP, “Audiobooks with virtual voice eligibility and troubleshooting” (invite-only beta, languages, US marketplace, pronunciation) (accessed November 10, 2026)
  15. ACX, “Learn about ACX” (open to residents of the United States, United Kingdom, Canada and Ireland) (accessed November 10, 2026)
  16. ACX, “ACX audio submission requirements” (human narration unless authorized, MP3 specs, levels, one chapter per file) (accessed November 10, 2026)
  17. Audacity, official website (free, open source audio editor for Windows, macOS and Linux) (accessed November 10, 2026)
  18. Audacity Manual, “Loudness Normalization” (LUFS or RMS target) (accessed November 10, 2026)
  19. EBU, Recommendation R 128, “Loudness normalisation and permitted maximum level of audio signals”, version 5.0 (November 2023) (accessed November 10, 2026)
  20. Lektify, documentation “Create your Lekty'Book” (importing a PDF or Word manuscript) (accessed November 10, 2026)
  21. Lektify, documentation “The Lekty'Book editor” (generating a chapter's audio narration) (accessed November 10, 2026)
  22. Lektify, documentation “Your Brand Kit” (“My voice”, your cloned voice) (accessed November 10, 2026)
  23. Lektify, documentation “Read in the Reader” (listening to the chapter) (accessed November 10, 2026)
Sonia Bé

Founder of Lektify