How to Convert a Dabble Manuscript Into an Audiobook With AI

If you write your novels in Dabble, your manuscript already lives in a clean, well-organized place. Turning that manuscript into a finished audiobook used to mean booking a narrator, renting studio time, or wrestling with recording software. With AudioProducer, you can take the same Dabble project and produce a narrated MP3 yourself, using AI voices, in an afternoon. This guide walks through the whole path: exporting cleanly out of Dabble, importing into AudioProducer, casting voices, reviewing the result, and exporting the finished file you then publish wherever you already sell your books.

Exporting your manuscript from Dabble

Dabble keeps your book in the cloud and gives you several export formats. For audiobook production, the two that matter are DOCX and EPUB. Open your project, go to the export option in the project menu, and choose one of these:

• DOCX (Word document): the most reliable choice. It preserves your chapter breaks and paragraph structure, and it opens in any word processor if you want to skim it before importing. The path from here is the same one you would follow to convert a Word manuscript to an audiobook.
• EPUB: also works well and keeps chapter divisions intact, which is handy if you plan to narrate chapter by chapter.

A few minutes of cleanup before you export pays off. Delete front-matter and back-matter that you do not want read aloud, such as the copyright page, the dedication if you prefer to skip it, and any author notes. Spell out anything that should be spoken in full, since an AI narrator reads text literally: "Chapter 12" reads better than "Ch. 12," and a number written as "1998" will be read as a year while "1,998" reads as a count. This kind of light pass keeps the narration from stumbling on abbreviations.

Importing your book into AudioProducer

Create a free AudioProducer account (no card required, and your first 1,200 words are on us so you can hear the quality before you commit). Start a new project and upload the DOCX or EPUB you just exported from Dabble. The text loads into the editor, where you can read through it, fix any leftover formatting, and split it into the sections you want narrated. If your book has natural chapter breaks, keep them: shorter sections are easier to review and re-generate if you want to change a voice or a line later.

The workflow is the same whether your draft started in Dabble, in Scrivener, in Vellum, or in Atticus: once the text is imported, the steps are identical. This is also the moment to decide how much you want to produce in one pass. You can generate a single chapter to test your voice choices first, then run the rest of the book once you are happy. Working in sections keeps the review manageable and lets you catch a mispronunciation early instead of at the end of a full novel.

Casting voices per character

A novel is more than a single narrator reading in a flat tone. AudioProducer lets you assign voices so your book sounds cast rather than dictated. You can pick one narrator voice for the whole book, or, for dialogue-heavy fiction, assign distinct voices to individual characters so a conversation actually sounds like two people talking. Browse the voice library, preview options against a paragraph of your own text, and choose the ones that fit your tone, whether that is a warm literary narrator or a brisk voice for a thriller.

Consistency matters across a full novel, so once you settle on a narrator and a cast, keep those assignments the same from the first chapter to the last. Listeners notice when a point-of-view character suddenly sounds different, and matching each voice to the same character throughout is what makes the finished book feel produced rather than pieced together. If you narrate in sections, note which voice belongs to which character so a chapter you generate next week matches the one you generated today.

If you want the book read in your own voice, voice cloning is available, with one firm rule: you may only clone a voice you have permission to use, which means your own voice or one you have explicit consent for. That consent requirement is there to protect writers and narrators alike.

Reviewing and exporting the MP3

Once your voices are set, generate the audio and listen back. Read along with your text and note anything that lands wrong: an unusual name, an invented place, a piece of dialogue that wants a different pace. When you find one, adjust the text (spelling a tricky name the way it sounds often fixes it) and re-generate just that section. You are not stuck re-recording an entire chapter to fix one word.

When the audio sounds the way you want, export it. AudioProducer produces a finished MP3 that you download to your computer. That downloaded file is yours to keep and to use however your publishing plan requires.

Where you publish the finished file

Here is the part worth being clear about: AudioProducer produces the audio file. It does not distribute or host your audiobook for you. There is no button that pushes your book to a store. What you get is the finished MP3, and you take that file and upload it wherever you already publish. If you sell direct from your own website, you host it there. If you distribute through an audiobook platform or aggregator, you upload the file into whichever service you already use. The production step and the distribution step stay separate, which means you keep full control of where your book goes and who carries it.

Frequently asked questions

See the FAQ below for the quick answers writers ask most often.