Can You Publish an AI Audiobook on Audible? (2026)

You cannot publish an AI audiobook you made yourself on Audible: ACX, Audible's self-publishing platform, says a submitted audiobook "must be narrated by a human unless otherwise authorized". The authorized AI routes onto Audible are Amazon's own, Virtual Voice for KDP ebooks and AI narration Audible produces for selected publishers. AI narration made with VoiceSmiths or any other tool can still be sold on Spotify, Google Play Books, Kobo and your own site, with disclosure. Below are each store's rules with sources, the audio specs, and what a VoiceSmiths render needs before upload. Rules were checked on October 26, 2026 and change often, so confirm on each store's page before you publish.

Store by store: who takes AI narration

So for a book you narrated with AI, the realistic plan is Spotify for Authors + Google Play Books + Kobo + direct sales, and, if you want Audible too, Virtual Voice (if your KDP ebook is invited) or a human narrator through ACX. On YouTube, also mind the monetization policies: AI content "made with generic or unoriginal templates giving the impression of mass production" is not monetized.

The audio specs stores ask for

Sources: ACX, Spotify for Authors metadata and assets guide (last updated November 2024), Google Play book file guidelines. ACX's specs only matter for a human-narrated ACX title, but they are a common reference, and Spotify's are close to them.

The render column is our own measurement on four clips from our pages, two on Eleven v4 and two on an older model, so treat it as a typical range, not a guarantee. The short version: against Spotify's and ACX's specs, a raw render misses on bit rate, peaks and head and tail silence, and its loudness varies by voice.

How to make a VoiceSmiths render store-ready

1. Render one chapter per file. A full cast takes up to 8,000 characters per paste (a longer chapter goes in as parts of one series); one narrator in long-form mode takes up to 50,000.
2. Decide where the AI notice goes. Full-cast renders carry an AI-generated label in their metadata and, by default, end with a spoken "This audio was generated by AI"; you can switch that off before rendering and disclose once in the opening credits and the store metadata instead. Single-narrator long-form files carry no label, so the disclosure is up to you.
3. Master the loudness. In Auphonic, choose the "RMS Levels (Audible / ACX)" normalization with a −20 dB RMS target and a −3 dB maximum peak. In Audacity, use Loudness Normalization and a limiter, then check the result with the ACX Check analyzer from Audacity's audiobook mastering guide.
4. Add the silence. Pad each file to 0.5–1 second at the start and 1–5 seconds at the end for Spotify.
5. Export 192 kbps constant-bitrate MP3, 44.1 kHz, mono. Re-encoding cannot restore detail lost in the original 128 kbps encode, but it meets the spec. For Google Play Books the 128 kbps mono file already meets the minimum.
6. Add the credits. Generate opening credits (title, author, narrator) and closing credits as their own short files, plus a retail sample where the store asks for one (ACX's limit is five minutes). Credit the narration truthfully (for example, "narrated by a digital voice").
7. Fill in the metadata. Tick the digital narration option on Spotify, choose Synthesized voice on Google Play, and name the AI voice as narrator on Kobo.

Rights and disclosure

• Audio rights first. You can sell an AI audiobook of your own book, of a book whose audio rights you have licensed, or of a public-domain text. Google Play, for one, makes owning the audio rights a condition of auto-narration.
• The audio is yours. Audio you generate on VoiceSmiths is yours to use commercially under section 6 of the terms of service.
• No borrowed voices. Our 24 character voices were designed from text descriptions, not cloned from anyone. If you clone a voice, use your own or one you have written permission for, and say so in the credits.

Hear what an AI audiobook chapter sounds like

A Christmas Carol, Stave One, from Dickens' text with a separate voice for the narrator, Scrooge and his nephew:

Read along in the theater →

The theater has 170+ English episodes made the same way. To produce your own, see the AI audiobook maker for one narrator or a full cast, novel to audiobook for a whole book step by step, and our AI audiobook production guide for the manuscript-to-release workflow.

What a whole audiobook costs to make

Both one narrator and a full cast cost 70 credits per 1,000 characters. A 90,000-word novel fits one 35,000-credit one-time pack: $749.90 at list price, or $299.90 at the new-user price in the first 24 hours after sign-up, and its credits never expire. Start with the trial pack (150 credits, about 2,100 characters; $3.90 new-user, list $9.90) to hear the voices on your own text. A first order is fully refunded within 7 days if you have used 150 credits or fewer. See pricing.

FAQ

Can you publish an AI audiobook on Audible? Not one you made with an outside tool. ACX, Audible's self-publishing platform, requires that a submitted audiobook be narrated by a human unless Audible authorizes otherwise. The AI routes onto Audible are Amazon's own: Virtual Voice, which turns an eligible KDP ebook into an audiobook (an invite-only beta for the U.S. marketplace), and AI narration that Audible produces with selected publishers.

Where can I sell an AI-narrated audiobook? Spotify, through Spotify for Authors with the digital voice option ticked; Google Play Books, with the narrator set to Synthesized voice; Kobo, which has said it accepts AI-narrated audiobooks with disclosure; and your own website. Apple Books takes only its own digital narration, and Voices by INaudio (formerly Findaway Voices) takes only unmodified packages from Google Play Books, ElevenLabs and Spoken Press. Check each store's current rules before you upload.

Do I have to disclose that an audiobook is AI-narrated? On the stores that accept it, yes. Spotify adds a sentence to the description when you tick "This audiobook uses digital voice narration", Google Play Books asks you to choose Synthesized voice as the narrator, and Virtual Voice titles are labeled by Amazon. Credit the narration honestly in your opening credits too.

Are VoiceSmiths audio files ready to upload to audiobook stores? Not without mastering. Renders are 128 kbps constant-bitrate mono MP3 at 44.1 kHz. That meets Google Play Books' minimum for mono, but Spotify for Authors asks for 192 kbps or higher, plus a loudness range and head and tail silence that raw renders do not reliably meet. Run each chapter through Auphonic or Audacity and export a 192 kbps MP3.

Can I publish an AI audiobook of a book I did not write? Only if you hold the audio rights: your own book, a book whose audio rights you have licensed, or a public-domain text. Note that Virtual Voice excludes public-domain titles, so that route is for books you publish on KDP yourself.

Open the studio and cast your first chapter →