Narrator: Audiobook MakerHow to make an audiobook for free
Get the book as an ePub, PDF or Word file, open it in a text-to-speech app on your iPhone, pick a voice and press play. Free tiers exist but all of them cap how much they’ll narrate per day, so a full novel usually takes weeks free or a few dollars to finish at once.
What you need
Three things, and only the first one is ever a real obstacle.
- A file you are allowed to open. An ebook or document with no DRM on it. The extension matters less than you would think — this is where most attempts stop, and the next section explains why.
- An app that reads it aloud. iOS has one built in. Dedicated apps do it better, mostly by handling chapters and remembering where you were.
- A decision about the voice. The gap between a modern AI voice and a 2015-era system voice is the difference between listening for two hours and giving up after ten minutes.
You do not need a computer, a converter, or an audio editor. Everything below happens on the phone.
Step 1 — Get your book as a file
Whether this works depends entirely on where the book came from, and it is worth being blunt about it because no amount of app choice changes the answer.
What will not work: books bought from the Kindle store or Apple Books. Both are DRM-protected, which means the file is encrypted so that only that company’s own app can open it. This is about the encryption, not the file type — an app can support .azw3 perfectly well and still be unable to open your Kindle purchase. Stripping the DRM is against the terms you agreed to when you bought the book. If you own a book on Kindle and want to hear it, your legal options are the Audible edition or Kindle’s own read-aloud where the publisher has enabled it.
What will work:
- Project Gutenberg and similar archives — around 70,000 public-domain books as free ePub or MOBI downloads, no account needed.
- DRM-free stores — Smashwords, itch.io, Humble Bundle book bundles, most technical publishers, and many self-published authors’ own sites.
- Your own documents — drafts, manuscripts, course notes, reports. Anything you exported from Word, Pages or Google Docs.
- Fanfiction archives — AO3 exports works as ePub directly from the download menu on any fic.
- Library ebooks — usually DRM-locked through Libby or OverDrive, so no. But many libraries also lend audiobooks outright, which solves the problem differently.
Once you have the file, save it into the iOS Files app. That is the common ground every reading app draws from — iCloud Drive, Dropbox, Google Drive, an email attachment or an AirDrop from your computer all end up there.
Step 2 — Import it
You have two broad options on an iPhone, and they are genuinely different tools.
The built-in one. iOS has Spoken Content (Settings → Accessibility → Spoken Content → Speak Screen). Turn it on, open a document, swipe down with two fingers, and iOS reads what is on screen. It is free, uncapped and already installed. It is also awkward for books: it reads the visible screen rather than the work, it loses your place, it cannot export anything, and it stops when the phone locks — which is exactly when most people want to listen.
A dedicated reading app. Apps built for this import the file properly, keep the chapter structure, remember your position, keep playing with the screen off, and usually let you export an audio file. Speechify, NaturalReader, ElevenReader, Voice Dream Reader and Narrator — the app behind this site — all work roughly this way, and all differ mainly in voice quality and where they put the paywall.
Whichever you use, the import itself is the same: open the app, add a book, pick the file out of Files. On a modern iPhone a 90,000-word ePub is ready in a few seconds.
Step 3 — Pick a voice
Do this before you start the book, not four chapters in. Play a paragraph or two of actual prose from your actual file — not the app’s demo sentence, which is always chosen to sound good.
What to listen for:
- Whether sentence endings land or run into each other. Bad prosody is what makes long listening tiring.
- How it handles names — invented names, foreign names, honorifics. Most engines get these wrong and none of them let you fix it word by word.
- Whether the pace is adjustable in small steps. Anything from about 0.8x to 1.5x should be usable; a fixed-speed voice is not.
- Which languages actually have a modern AI voice. Apps often advertise a large language count where most of the list falls back to old system voices.
That last point is worth checking against whatever app you pick. In Narrator, for instance, the natural AI voices cover English, French, Spanish, Italian, Portuguese, and the other 25+ languages use the voices built into iOS. Both are real; they are not the same experience.
Step 4 — Listen, or export a file
For most people listening in the app is the whole point: press play, lock the phone, put it in a pocket. Check that playback survives the screen going off and that the lock-screen controls work — this is where the built-in iOS option falls down.
Exporting is the other route. An app that can write out an .m4a file gives you something that behaves like any other audio: it plays in any app, survives you deleting the source document, and can be moved to another device. It is also slower, because the whole book has to be narrated up front rather than as you go.
What "free" actually gets you
This is the part every vendor page skips, and it is the reason people add “reddit” to these searches. So, honestly:
Generating a natural-sounding voice costs the app money for every character it reads. That is why the pattern is so consistent — the apps with genuinely good voices cap the free tier, and the apps with no cap use the old system voices. There is no app that gives you unlimited free listening in a modern AI voice, and any page telling you otherwise is describing a trial.
The caps take different shapes: a character limit per day, a minutes-per-month allowance, a limited number of imported files, or a free tier that only reads web pages and not your own documents. That last one catches a lot of people — the app works beautifully until the moment you import your own book.
Narrator’s version of this, since you are on its site:
Free plan at a glance
No account, no trial expiry. The cap is on narration — that means listening, not just export.
| Every day | 25,000 characters — about 4,000 words, roughly one chapter, or 28 minutes of listening at 1x. Resets daily. |
|---|---|
| Lifetime pool | 150,000 characters — about 25,000 words, or 2.8 extra hours. Only drawn on after you pass the daily limit. |
| A full novel | An average 90,000-word book is about 22 days of free listening. A 400,000-word one is closer to 96. |
| Plus | $4.99/month or $29.99/year. Removes the limits. |
Which is to say: about a chapter a day, indefinitely, for nothing — an average 90,000-word novel in roughly 22 days. If you want to finish a book this weekend rather than next month, that costs $4.99. We would rather write that down than have you find out on day two.
Common problems
Four things go wrong, and they go wrong in roughly this order of frequency.
- The PDF reads nothing. It is almost certainly a scan — an image of a page rather than text. Test it by trying to select a word in Preview or Books. If you cannot select it, no app can read it without running OCR over it first.
- The chapters are in the wrong order, or there are none. ePubs carry a table of contents; PDFs and plain text do not. Converting a PDF to ePub first, with a desktop tool like Calibre, usually fixes both the ordering and the ability to skip around.
- It reads the front matter forever. Copyright pages, dedications and long tables of contents get narrated like anything else. Skip to the first real chapter manually the first time.
- A very large file is slow or refuses to import. Omnibus editions and 400,000-word fics are genuinely big. Splitting the file into parts is more reliable than waiting.
What it doesn’t do
- DRM-protected books from Kindle and Apple Books cannot be converted by any app.
- Scanned PDFs have no text layer and will not read.
- Synthetic narration cannot be sold or published — it is for your own listening.
- Free tiers with good voices are always capped somehow. Check where the cap is before you start a long book.
Common questions
Not if you bought it from the Kindle store. Plenty of apps read Amazon’s file formats — MOBI, AZW and AZW3 — but a store purchase is encrypted on top of the format, and that encryption is what stops another app opening it. Some Kindle titles have Audible narration you can buy, and the Kindle app has a basic read-aloud on some titles, but there is no legal way to turn a DRM-protected file into your own audio file.
The one genuinely uncapped option is Apple's built-in Spoken Content, because it is part of iOS rather than a service someone pays to run. The trade-off is that it reads what is on screen, cannot export a file, and the voice quality and pacing controls are limited. Every app with good AI voices caps its free tier somehow — by characters, minutes or files.
At an audiobook's usual pace, a 90,000-word novel is roughly 10 hours of listening. On a free tier capped at 25,000 characters a day it works out at about 22 days, a chapter at a time.
No. Synthetic narration of a book you did not write is for your own listening. Retailers like Audible and ACX have their own rules about AI narration and about rights, and converting a copyrighted book you bought does not give you the right to distribute it.
If you want to try it on this phone
Narrator does the four steps above: it imports ePub, MOBI, AZW3, PDF, DOCX, RTF and TXT, narrates with AI voices in five languages, keeps playing with the screen locked, and exports an .m4a file. The free plan is the 25,000 characters a day set out above, with no account and no trial expiry.
Free plan, no account · 25,000 characters a day · iOS 26+
Scan to open on your iPhone