How to turn your EPUB into an audiobook (step by step)

Updated July 2026

If you've published an ebook, you already own the hardest input to an audiobook: a clean, chaptered EPUB. What stands between that file and a listing on Spotify or Apple Books is narration, audio mastering, and packaging. There are three honest ways to get there.

Option 1 — Human narration ($1,200–$2,800 per book)

A professional narrator typically charges $200–$400 per finished hour, and a 90,000-word novel runs about 9–10 finished hours. Add studio-quality proofing and you land in the $1,200–$2,800 range, with 4–8 weeks of turnaround. For a strong frontlist title with proven sales, this is still the premium option — and it's the only way onto Audible via ACX. For a backlist of five books, it's $10,000+ that most indie authors will never recover.

Option 2 — DIY with AI voice tools

Text-to-speech tools (ElevenLabs and similar) can generate excellent narration, but "generate speech" is maybe a third of the job. Doing it yourself means: splitting your EPUB into chapters, feeding the text through in chunks, stitching thousands of audio segments, normalizing loudness to retail specs (roughly −18 to −23 LUFS, −3 dB peak), building a chaptered M4B with cover art, and writing the AI-narration disclosure each store requires. Budget a weekend or two of technical work per book, plus subscription costs — often $99+ for enough characters for one novel.

Option 3 — Done-for-you conversion (from $49)

Purpose-built converters do the whole pipeline. The workflow with Bookvox:

Upload your EPUB — or your Word manuscript. Both .epub and .docx work; chapters are detected automatically (heading styles or "Chapter One"-style lines), and front matter (copyright pages, table of contents) is excluded from narration by default. Writing in Word and don't have an EPUB at all? See the Word-document guide.
Pick a narrator and listen free. A full sample chapter costs nothing — judge the voice on your actual prose, not a demo clip.
Fix pronunciations. Character and place names get a pronunciation dictionary (e.g. "Naoise → NEE-sha") applied to every chapter.
Pay once, render everything. One payment by book length ($49 up to 50k words, $79 to 100k, $119 beyond). Every chapter is narrated and mastered to retail loudness.
Download the retail package. Mastered per-chapter MP3s, a chaptered M4B with your cover embedded, and the disclosure text — ready for Spotify, Apple Books, Google Play, and Kobo (see the platform rules — Audible is the one holdout).

The sample-edition mistake. The most common conversion failure we see: uploading a retailer preview EPUB instead of the full book. Store downloads are often samples that end mid-chapter — sometimes only 8,000 words of a 100,000-word book. Always export the complete EPUB from your writing tool (Vellum, Atticus, Scrivener, Calibre, or your KDP source file), and check the word count shown after upload against what you know your book to be.

What "retail-ready" actually means

Loudness-mastered audio — stores expect consistent levels; unmastered TTS output gets rejected or sounds amateur next to other titles.
Chapter markers — listeners expect to skip by chapter; that requires a proper M4B or correctly split MP3s, not one giant file.
Embedded cover art and metadata matching your ebook listing.
AI disclosure — every major store that accepts AI narration asks you to declare it.

Hear your book before you spend anything.

Upload your EPUB and get a full sample chapter free — no signup. Start with your book →