How to Convert an EPUB to an Audiobook

Most books never get an audiobook release. If the book you want to listen to exists as an EPUB, you have three realistic ways to turn it into one: a hosted converter, an open-source tool you run yourself, or the text-to-speech built into your phone or e-reader. This guide covers all three honestly; they suit different people.

First, the ground rules

Every method below needs a DRM-free EPUB, meaning a file whose text the tool can actually read. Books you wrote, public-domain books (Project Gutenberg, Standard Ebooks), DRM-free stores, and EPUB exports from sites like Royal Road all qualify. Kindle and Kobo purchases are generally DRM-locked and won't work. Convert only books you own or have the right to convert, for your own listening.

Method 1: A hosted converter (the fast path)

A hosted converter does the whole pipeline in the browser: text extraction, narration with a neural voice, and packaging into an audiobook file. This is the right method if you want the result in minutes and don't want to install anything.

With LibriTalk (disclosure: that's us), the flow is:

1. Upload the EPUB. You get a free preview showing the narrative word count and which non-narrative sections (table of contents, copyright page, index) were excluded.
2. Pick one of 10 AI voices: US or British, male or female, each with a playable sample.
3. Wait roughly 15 minutes for a mid-length book.
4. Download an M4B with chapter markers that match the book's real chapters.

The first book under 100,000 words is free with no card; after that it's $0.99–$4.99 per book depending on length. The honest limitation: AI narration is a single consistent voice, clear and natural, but it won't act out characters the way a professional narrator does.

Method 2: Open-source tools (the tinkerer's path)

If you're comfortable with Docker or Python and have decent hardware, ebook2audiobook is an excellent open-source project that converts EPUBs into chaptered audiobooks using local TTS engines such as XTTSv2, Bark, and VITS. It costs nothing per book, supports voice cloning with some engines, claims over 1,100 languages, and runs entirely on your machine. Free hosted demos also exist on Hugging Face and Google Colab, subject to the usual free-tier quotas and timeouts.

The trade-offs are time and setup: installation, model downloads, and conversion that can take hours per book on a CPU (a GPU helps a lot). For converting an entire library at zero marginal cost, it's the rational choice. We wrote a detailed LibriTalk vs ebook2audiobook comparison if you're weighing the two.

Method 3: Built-in text-to-speech (the free-right-now path)

Your devices can already read EPUBs aloud, with caveats:

• iPhone/iPad: open the book in Apple Books and use Spoken Content (Settings → Accessibility) to have the system voice read the screen.
• Android: Google Play Books and Moon+ Reader can read aloud with system TTS voices.
• Desktop: Calibre's viewer has a read-aloud function, and most e-readers have some TTS option.

This costs nothing and works today. What you don't get is an audiobook: there's no file, no chapters, the screen usually has to stay on the book, and the voices are a tier below modern neural TTS. It's listening-while-reading rather than a replacement for an audiobook.

Which method should you pick?

• You want an audiobook today with zero setup → hosted converter.
• You're technical and converting many books → ebook2audiobook on your own hardware.
• You just want this one chapter read to you right now → built-in TTS.

Whichever route you take, insist on M4B output with chapter markers for anything book-length, because a chapterless multi-hour MP3 is genuinely unpleasant to navigate. If you're not sure what M4B is, read What is an M4B file?, and once you have one, see how to play it on iPhone and Android.