Turning a finished book into an audiobook is not simply a matter of pressing “generate.” You need the audio rights, a narration workflow that fits your manuscript, careful listening and correction, accurate store metadata, and a distribution route that accepts the kind of narration you plan to use.
Most importantly, platform rules are not interchangeable. One retailer may offer its own digital narration, another may allow AI narration only through a controlled program, and another may require human narration unless it has expressly authorized an exception. Check the current requirements for each destination before committing to production.
1. Confirm rights before you choose a voice
Start with the rights agreement for the book. You need the right to make and distribute an audiobook in the territories and languages you intend to serve. Owning the print or ebook rights does not automatically prove you have audio rights; read your publishing contract, co-author agreement, translation license, and any other relevant contracts. Google, for example, says creators must own audio rights for titles selected for its auto-narration program, while Apple also requires audio rights for its program.
If you use a synthetic or cloned voice, check the narration provider’s current license and the distributor’s rules separately. A tool’s commercial-use permission does not itself authorize copying a real person’s voice or guarantee retailer acceptance. Get documented permission before using an identifiable person’s voice replica. Laws vary by jurisdiction; the U.S. Copyright Office describes a patchwork of existing protections for digital replicas. Seek qualified legal advice for a disputed rights question.
2. Prepare the manuscript for listening
Work from the final, proofread edition. Audio makes awkward sentences, inconsistent names, typos, and missing transitions more obvious. Save a clean source file and narration copy; remove artifacts such as page numbers, running headers, footnote markers, and ebook navigation text. Preserve headings and chapter order for navigation.
Plan how to handle tables, charts, image captions, endnotes, URLs, abbreviations, equations, and lists. They may need spoken explanations or an audio-friendly rewrite, not silent deletion. Apple says its digital narration is optimized for narrative, reflowable English ebooks; complex formatting and visual-heavy works may not be suitable.
Make a pronunciation sheet for names, acronyms, foreign phrases, and specialized terms. Mark speaker changes and meaningful pauses. Tools differ: Spotify describes pronunciation rules, while Google offers controls for voice, pronunciation, pacing, and editorial fixes.
3. Pick a production path that allows your narration
Choose among self-produced audio, a retailer’s digital-narration workflow, human narration, or a distributor that delivers finished audio to several stores. Compare territory, rights and exclusivity, royalty and pricing terms, file requirements, reporting, and takedowns. One distributor’s acceptance does not guarantee acceptance by every retailer.
Platform rules illustrate why this check matters:
Apple Books: Apple’s digital narration program currently takes nominations through preferred partners. Its listed eligibility includes an English-language ebook already available on Apple Books, audio rights, and a reflowable format; Apple reviews each title and may reject it even if it meets the basic criteria. Apple states the result is labeled “Narrated by Apple Books.” Its partner guidance also requires the audiobook description to begin: “This is an Apple Books audiobook narrated by a digital voice based on a human narrator.” The program is not an open upload of any AI-narrated audio file.
Spotify for Authors: Spotify’s September 2026 announcement says its built-in Audiobook Creation Tools beta is available to self-published authors in the United States for English-language books, currently up to five titles without charge; a validated payment profile and other usage limits apply. It accepts EPUB manuscripts, allows preview and approval before publication, labels digital-voice titles for listeners, and is non-exclusive. Availability and limits are specific to this program and may change. Authors can also publish finished audiobooks directly through Spotify for Authors, and Spotify describes a separate opt-in route to wider retail distribution via Voices by INaudio.
Google Play Books: Google’s auto-narration program creates audio from an eligible ebook, requires audiobook rights, and provides an audio download. Its policies currently say an auto-narrated edition sold elsewhere must also be on Google Play in every country where it is sold elsewhere, and that the Play list price cannot be higher than the equivalent elsewhere. Review the program’s current language and account-availability rules rather than assuming every ebook qualifies.
Audible/ACX: ACX’s current submission guidance says a submitted audiobook must be human-narrated unless otherwise authorized; unauthorized text-to-speech, AI, or automated recordings are prohibited. ACX has separately described a limited U.S. beta involving participating narrators’ approved voice replicas. That controlled exception is not blanket permission to upload any AI narration.
If the same title goes to several retailers, verify that your contracts and chosen production route permit each delivery. In particular, a retailer’s built-in narration program may have conditions that differ from the rules for independently produced audio.
4. Generate, edit, and listen all the way through
Use a consistent workflow: generate by chapter or logical section, then review each rendered file against the manuscript. Listen on headphones and on an ordinary phone or speaker. Correct misread words, skipped lines, odd pauses, repeated text, swallowed sentence endings, distracting changes in loudness, and voice shifts. Names that sound fine in a short sample can become tiring or confusing across a whole book.
Review the opening and ending as carefully as the chapters. Include credits and narrator information where the selected platform requires them. Confirm that chapter titles are spoken or represented in the correct files, that transitions do not cut off words, and that every section is in the right order. If you change the manuscript after generation, regenerate and recheck the affected audio rather than assuming the old files still match.
Technical delivery targets depend on the retailer. For example, ACX specifies one chapter or section per file, separate opening and closing credits, and MP3 files at 44.1 kHz with at least 192 kbps constant bit rate. Its current guidance also sets RMS, peak, and noise-floor limits. These are ACX-specific requirements, not a universal audiobook standard—and they apply only if the title is eligible for ACX submission in the first place. Google lists different accepted formats and minimum technical thresholds for its own audiobook uploads. Use the destination’s latest checklist and run its available validator before delivery.
5. Finish the metadata and cover
Prepare one accurate record of the title, subtitle, author, contributors, narrator or digital-narration disclosure, language, series, publisher, description, release date, rights territories, and identifier. Keep it consistent across cover, credits, file names, and retailer forms; do not stuff keywords into the title.
Choose a square audiobook cover where the retailer requests or recommends one, and check the dimensions and file type for each destination. Apple allows custom audiobook art that meets its specifications or can adapt the ebook cover; Google publishes its own pixel and format requirements. Do not assume an ebook cover file automatically meets every audio-store specification.
An audiobook is an ISBN-eligible product under the International ISBN Agency’s guidance, but retailers may provide or require different identifiers. Check your national ISBN agency and each platform’s instructions before buying or assigning one. If you use an ISBN, make sure it identifies the correct audiobook edition and is recorded consistently.
6. Publish, verify the listing, and keep records
Complete rights and content declarations, upload final files and metadata, choose territories and a release date, and review distribution terms. Monitor the dashboard for corrections. Apple says digital narration may take one to two months and titles undergo quality review, so allow for uncertainty in the launch date.
Once a store listing is live, check the public page on desktop and mobile. Confirm the title, author, narrator disclosure, cover, description, sample, chapters, and purchase or playback behavior. Save the approved master audio, metadata, cover source, rights documentation, delivery receipts, and retailer reports. If you later replace files or unpublish, record the date and destinations affected.
For an AI-assisted workflow, see the
FlashTTS audiobook page for its current availability and early-access signup. Treat generated files as drafts until reviewed and accepted by your chosen retailer.
FAQs
Can I upload any AI-narrated audiobook to Audible?
No. ACX requires human narration unless otherwise authorized; its voice-replica beta is limited and controlled. Check current rules first.
Can I distribute one AI audiobook to multiple stores?
Sometimes. Spotify describes its creation tools as non-exclusive, while Google’s policy adds conditions for selling the same auto-narrated edition elsewhere. Check each destination.
Do I need to disclose AI narration?
Follow each destination’s instructions. Spotify labels digitally voiced titles, while Apple identifies its titles as narrated by Apple Books and specifies description wording.
Does AI narration mean no editing is needed?
No. You remain responsible for the manuscript, pronunciation, completeness, pacing, audio consistency, metadata, and rights. Listen to the full audiobook before publishing; an automatic voice can still misread, skip, or awkwardly render text.