🎧 Listen to this article
On This Page
0%You finished the manuscript, maybe even the ebook is live on KDP. Now you want the audiobook everywhere — not locked into one store on a seven-year exclusive deal. "Going wide" means distributing your audiobook to Audible, Spotify, Google Play Books, Apple Books, Kobo, and library systems at once, on non-exclusive terms. The blocker used to be production cost. A human narrator runs $200-$400 per finished hour, and that math rarely pencils out before you've sold a single copy.
This tutorial walks you through producing a wide-ready audiobook with AI narration and pushing it to the major retailers — start to finish, with the file specs each platform actually checks.
TL;DR — the wide audiobook workflow in 6 steps:
- Decide wide vs. exclusive (royalty rate vs. reach trade-off).
- Prep your manuscript: clean text, pronunciation list, chapter breaks.
- Narrate each chapter with a consistent AI voice.
- Master to retail spec: -23 to -18 LUFS, -3 dB peak, mono or stereo, MP3 192kbps CBR or higher.
- Add disclosure metadata where the platform requires it.
- Upload through a wide aggregator (Findaway Voices by Spotify, Author's Republic) plus Audible direct.
Should you go wide or stay exclusive?
This is the first real decision, and it's about royalty rate versus reach. Exclusive distribution through ACX pays a higher per-unit royalty (40% on Audible) but locks the title to that ecosystem for the term. Wide/non-exclusive pays less per unit (25% on ACX) but lets the same file sell on Spotify, Google Play, Apple, Kobo, and library platforms like Libro.fm and OverDrive.
| Factor | Exclusive (ACX) | Wide (non-exclusive) |
|---|---|---|
| Audible royalty | 40% | 25% |
| Spotify / Google Play / Apple | No | Yes |
| Library distribution | No | Yes (OverDrive, Libro.fm) |
| Term commitment | 7 years | None — leave anytime |
| Best for | One-store sellers | Catalog authors, cross-promo |
A practical rule: if your readers already live on Audible and you want the highest per-sale cut, exclusive can win on a single title. If you have a series, a newsletter, or a platform of your own to drive traffic, wide almost always wins on total revenue because you're not turning away every non-Audible listener — and Spotify Audiobooks has reportedly been one of the faster-growing channels going into 2026.
Step 1 — Prep your manuscript
Narration quality starts with clean input. Export your manuscript to plain text or a clean DOCX and strip anything the listener can't hear: page numbers, headers, footnotes that don't read aloud, image captions, and "see figure 3" references.
Build a short pronunciation list for proper nouns, invented words, and acronyms. Decide up front whether "Dr." reads as "Doctor," how character names are stressed, and whether numbers read as digits or words. Keeping this list in front of you means every chapter stays consistent.
Finally, mark your chapter breaks explicitly. Most retailers want one audio file per chapter, named in playback order (01-opening-credits, 02-chapter-one, and so on).
Step 2 — Narrate each chapter
Load each chapter into AudioPod's audiobook studio and generate the narration. Two things to get right here:
- Pick one voice and stay with it. Switching voices between chapters is the fastest way to sound amateurish. Audition a few on a representative paragraph — dialogue-heavy if it's fiction, dense if it's nonfiction — then commit.
- Use the pronunciation list. Feed your proper-noun corrections in before you render the full chapter, not after. Re-rendering a 40-minute chapter to fix one name is wasted time.
If you'd rather start from the finished book file directly, the AI narrator and audiobook generator take a manuscript and produce chaptered audio without you pasting text block by block. For a quick proof of concept before you commit a budget, the free audiobook generator lets you hear how a chapter reads.
Listen to the full render once, start to finish. AI narration is strong in 2026 — indistinguishable from a human read in most genres — but you still catch the occasional flat line reading or a homograph ("lead" the metal vs. "lead" the verb) that wants a pronunciation hint.
Step 3 — Master to retail spec
Every retailer runs an automated audio check on upload, and most rejections are loudness or peak failures, not content. Master each chapter to these targets, which satisfy ACX and the wide platforms at once:
| Spec | Target |
|---|---|
| Integrated loudness | -23 to -18 LUFS |
| Peak | no higher than -3 dB |
| Noise floor | -60 dB RMS or quieter |
| Format | MP3, 192 kbps CBR minimum (many wide stores accept 256) |
| Sample rate | 44.1 kHz |
| Channels | mono is fine; stereo accepted |
| Per-file structure | one file per chapter |
| Room tone | 0.5-1 s of silence at head and tail |
If a chapter comes back too quiet or too hot, normalize it and re-check rather than guessing. If you recorded any human intro or pickup and it has hiss, run it through noise reduction before mastering so your noise floor passes. Need to convert a stray WAV to a compliant MP3? The audio reader and converter tools handle the format step.
Step 4 — Add your AI-narration disclosure
Audible and ACX accept AI-narrated audiobooks in 2026 with disclosure. The disclosure lives in metadata, not in the audio. When you submit through ACX or a wide aggregator, you'll check a box or select a production-method field indicating the title uses a synthetic/AI voice. Google Play Books has its own auto-narration disclosure flow. Don't skip this — undisclosed AI narration is the one thing that reliably gets a title pulled.
You do not need to add a spoken "this audiobook was AI-narrated" line unless a specific store asks for it. Check each platform's current submission form, since these fields have shifted over the past year.
Step 5 — Upload to Audible direct
For the Audible/Amazon/iTunes leg, submit through ACX on the non-exclusive plan (25% royalty). Upload your chaptered MP3s, the retail and a 1-5 minute sample clip, your cover at 2400×2400 px square, and complete the metadata including the AI-narration disclosure. Approval review typically takes a couple of weeks.
Step 6 — Upload to a wide aggregator
For everywhere else, use a single aggregator so you're not creating ten store accounts. The two common choices:
- Findaway Voices by Spotify — pushes to Spotify Audiobooks, Apple, Google Play, Kobo, libraries, and more; per-title or à la carte.
- Author's Republic — similar reach, different fee structure.
Upload the same mastered chapter files and cover. Set your price (or let the platform set it), confirm territories, and re-enter the AI-narration disclosure. Because your files already pass ACX spec, they'll clear the wide platforms' checks too — the specs converge.
Common errors and fixes
| Symptom | Cause | Fix |
|---|---|---|
| Upload rejected: "audio too loud" | Peak above -3 dB | Re-normalize to -3 dB peak, re-check LUFS |
| Upload rejected: "too much background noise" | Noise floor above -60 dB | Run noise reduction, re-master |
| Voice sounds different in chapter 7 | Switched voice or settings mid-project | Re-render with the same voice profile |
| Name mispronounced throughout | No pronunciation hint before render | Add to pronunciation list, re-render that chapter |
| Title flagged after going live | AI narration not disclosed | Update production-method metadata in the dashboard |
Frequently asked questions
Can I use the same audio files for ACX and the wide platforms? Yes. If you master to the spec table above (-23 to -18 LUFS, -3 dB peak, 192 kbps+ MP3, one file per chapter), the same files clear both ACX and the wide aggregators. The standards have largely converged.
Do I have to disclose AI narration? Yes, where the platform asks. It's a metadata field on ACX, Google Play, and the wide aggregators — not a spoken line. Undisclosed AI narration is the main thing that gets titles pulled.
How long does my audiobook stay available on AudioPod after I make it? Generated files are retained on a sliding window by plan — 1 year on the free tier, 2 years on Creator, 3 on Pro, 5 on Studio — and opening or downloading a file in the last 30 days extends it. Your job history stays in the dashboard regardless, so you can re-render a chapter in one click. See pricing for the full table.
What does it cost to narrate a full book with AI versus a human? A human narrator commonly runs $200-$400 per finished hour, so a 10-hour book is $2,000-$4,000 before pickups. AI narration on a free or low-cost plan is a fraction of that, which is what makes going wide viable before you've made a sale.
Can I go wide and stay on Audible at the same time? Yes — that's the whole point of non-exclusive. Submit to ACX on the 25% non-exclusive plan and to a wide aggregator in parallel. You only lose the Audible 40% rate if you'd otherwise have gone exclusive.
Should I narrate fiction and nonfiction the same way? Not quite. Fiction benefits from auditioning voices on a dialogue-heavy passage to check character delivery; nonfiction reads better with a steady, clear voice on a dense paragraph. Either way, pick one voice per book.
Start with one chapter
The lowest-risk way to learn this workflow is to narrate a single chapter, master it to the spec table, and run it through one platform's upload check before you commit the whole book. The audiobook studio and the free audiobook generator both start on the free tier — 1,000 credits, every tool, no card. Produce a chapter, confirm it passes, then scale to the full title and push it wide.
