Manuscript to Audiobook for Self-Published Authors

Prepare your manuscript

A little preparation makes manuscript to audiobook faster and saves listening time later. Open each chapter with a Word heading style or a clear "Chapter 12" line, take out what shouldn't be read aloud (the copyright page, the also-by list, image captions, usually footnotes) or untick those sections when you import, and keep dialogue tags where a long exchange could lose the listener. The formatting rules, with a template, are in audiobook script format; if your manuscript lives in Word, see Word to audiobook.

Cast your book: narrator plus character voices

Pick the narrator first, because it carries most of the running time: a voice that is clear and easy to listen to for hours. Then give the main characters voices that contrast in age, pitch or accent, and save the most distinctive voices for smaller parts. Play a full chapter before committing; a voice that charms in one line can tire the ear over twenty minutes.

Choose standard voices (Google Gemini) or premium voices (ElevenLabs). Standard voices use fewer credits, which suits drafts and proof-listening, and you can switch the project to premium voices for the final production.

Proof-listen and fix lines before export

Listening is the most honest edit a manuscript gets. Your ear catches what your eyes skip: the same word three times in a paragraph, a sentence that runs out of breath, dialogue nobody would say out loud. Speaker detection doubles as a clarity check. If the studio gives a line to the wrong character, ask whether a reader could tell who was speaking. To hear drafts while you're still writing, see text to speech for writers.

Fixing is line by line. Change the speaker, edit the text, add a direction like [whisper, like sharing a secret], and regenerate just that line; regenerating uses credits again, so batch your fixes. Make the first chapter your sample: the free credits on sign-up cover a short sample chapter, and it's the right thing to share with beta readers, your newsletter, or an agent or publisher you're talking to about audio. Label it as AI narration when you do.

Self publish audiobook: where AI narration is allowed in 2026

If you self publish audiobook editions, check each store's current policy before you upload. As of October 2026: ACX, which supplies Audible, requires narration by a human unless otherwise authorized under its April 15, 2026 rules, though the same page says Audible is working to accept third-party text-to-speech. Spotify for Authors accepts digital-voice narration when you tick the box that labels it. Other stores and distributors set their own rules; where to sell AI audiobooks goes through them.

Amazon's own KDP Virtual Voice is an invite-only beta for KDP ebooks, distributed in the US marketplace only, with one voice per chapter and an audio file that Amazon keeps. If you want character voices and files you own, compare the two in KDP Virtual Voice alternative. Selling directly from your own website or newsletter is another route, where the rules are yours; selling audiobooks on your website covers the setup.

Costs, rights and owning your audio files

Credits are counted by the characters voiced, and the import screen estimates length and credits before you create the project. For a number up front, put your word count into the audiobook cost calculator, which shows the running time, the credits and which plan covers it.

Under our terms, you own the audio and exports you generate and may use them for any lawful purpose, including selling them. Two caveats: in some countries AI-generated audio may not be protected by copyright, or only in part, and library voices are open to every user, so someone else's audiobook may sound similar. See how the whole process fits together in how it works.