Voice engines, retailer tools and book creators that narrate: what each costs, and where the audio can go.
If cost and reach matter most, start with the retailers’ own tools. KDP Virtual Voice and Google Play’s auto-narration are free to produce with and put the title straight in their stores.
If you want the widest choice of voices and already have a finished manuscript, go to ElevenLabs directly. Anthra’s narration uses ElevenLabs voices, and ElevenLabs’ own plans cost less per character.
Choose Anthra if your book already lives in Anthra and you want narration in the same workflow: a voice per character, a pronunciation list, and only changed paragraphs re-narrated after edits. It makes most sense for picture books and shorter works, because narrating a novel is expensive.
ACX and Audible are not a route for audio made in Anthra. ACX requires human narration unless ACX has authorized otherwise, and that holds whatever the file quality.
As of November 2026, unless a date is given. Competitor and retailer details come from each company’s own website (linked below), describe features as that company does, and may have changed since. This page is research, not legal advice.
| Tool | Best for | Price | Voices | Where the audio can go |
|---|---|---|---|---|
| ElevenLabs | Best forVoice range and per-character casting from a finished manuscript | PricePlans from $6/mo (30k credits) to $990/mo (6M); about 1 credit per character. It says audiobooks run from free to about $200 per book | Voices10,000+ voices, one per character; it says it detects characters and matches voices to them | Where the audio can goDirect distribution to ElevenReader, Spotify and INaudio. ElevenReader pays 60% direct and $0.20 per hour streamed, with no exclusivity |
| KDP Virtual Voice | Best forReaching Amazon and Audible listeners with no production cost | PriceFree to produce; 40% royalty on à la carte sales; list prices $3.99 to $14.99 | Voices80 Amazon voices; a different voice per chapter | Where the audio can goAmazon, Audible, Alexa and Amazon Music Unlimited. KDP’s page describes a US beta with an interest list |
| Google Play auto-narration | Best forFree narration for titles sold on Google Play | PriceFree “for a limited time”; 52% royalty on Play sales | Voices50+ voices in 6 languages | Where the audio can goGoogle Play. The audio can be downloaded, and sold elsewhere only if it’s also on Play |
| Spotify Audiobook Creation Tools | Best forUS authors with English-language titles who want to be on Spotify | PriceSee Spotify’s page; limited to 5 projects per account | VoicesElevenLabs voices | Where the audio can goSpotify, with titles labeled “Narrated by Digital Voice” |
| Inkfluence AI | Best forLow-cost narration bundled with a writing subscription | PricePlans at $9.99 and $19.99/mo | VoicesIt says 49 voices in 14 languages, with a different narrator per chapter | Where the audio can goIt says it exports a ZIP of per-chapter MP3s at 44.1 kHz, 192 kbps CBR, mono: ACX’s technical spec, though ACX still refuses AI audio |
| Speechki | Best forMany languages, with roles assigned in a visual editor | PriceNo public pricing | VoicesIt says 1,100+ voices in 80+ languages | Where the audio can goNo public retailer list |
| Anthra | Best forNarrating a book you wrote or imported in Anthra, especially picture books and short works | Price4 credits per 1,000 characters, pay as you go, no subscription | VoicesElevenLabs voices, cast per character. Dialogue is matched to speakers by scanning for names, so listen through the result | Where the audio can goWhole-book M4B with chapter markers, an MP3 per chapter, a retail audio package, Anthra’s reader and a podcast feed. Not a route to ACX or Audible |
Voices. ElevenLabs says it offers 10,000+ voices, detects the characters in your book and matches a voice to each, and distributes directly to ElevenReader, Spotify and INaudio. Anthra’s narration runs on ElevenLabs voices, so for audio alone, going direct gives you more choice for less money.
Reach and price. KDP Virtual Voice puts a title on Amazon, Audible, Alexa and Amazon Music Unlimited at no production cost, and KDP says Virtual Voice titles are enrolled automatically in KDP Select when the ebook is, which brings them into Kindle+ in the US. Audio made in Anthra has no route into those channels. Google’s auto-narration is also free for now.
Retail files. Inkfluence AI says it exports per-chapter MP3s at 44.1 kHz and 192 kbps CBR, which is the technical spec ACX publishes, and its plans start at $9.99 a month. If you need files for a specific retailer, compare every tool’s output with that retailer’s current specs before you pay.
ACX and Audible: ACX’s requirements say audio must be human-narrated unless ACX has authorized otherwise, and that unauthorized use of text-to-speech, AI or automated recordings is prohibited. The page also says Audible is working to accept third-party TTS content, with no date. So this route is closed to audio made in Anthra.
Amazon: Virtual Voice uses Amazon’s own voices only.
Google Play: Google’s partner help accepts third-party AI audio marked “Synthesized voice”, and lists MP3, AAC, FLAC, WAV, ZIP and LPF files.
Spotify: Spotify’s digital voice page names providers “like Google Play Books, ElevenLabs, and more”, asks you to tick a disclosure, and says these titles aren’t shared with referral partners. We can’t confirm which other sources qualify.
Voices by INaudio: INaudio takes AI audio only as unmodified LPF files from Google Play Books, ElevenLabs or Spoken Press. Anthra isn’t one of those three.
Anthra makes no promise that any retailer will accept audio made in it. Read the retailer’s current rules and file specs before you upload, and disclose AI narration where it asks.
Anthra narrates the same manuscript you write or import, so there’s no separate upload. You cast a narrator and a voice per character. Dialogue is matched to speakers by scanning the surrounding text for known names, and anything it can’t attribute falls back to the narrator, so listen through the result. A per-book pronunciation list handles invented names, and after edits only the changed paragraphs are narrated again.
You get a whole-book M4B with chapter markers, cover art, optional ambience beds and intro and outro stings, plus an MP3 for each chapter. For store upload there is a separate retail audio package: one MP3 per chapter at 192 kbps CBR, 44.1 kHz mono, plus opening and closing credits files and a measured loudness report. That is a file spec, not a retailer’s approval. Listeners can also use Anthra’s reader or a per-book podcast feed. If you sell the book on Discover, readers pay with Anthra credits, not at a card checkout.
We don’t claim these files are mastered to a retail spec. They’re made for listening, so check them against your retailer’s specs if you plan to upload them anywhere.
Anthra charges 4 credits per 1,000 characters, rounded up per chapter. A 1,000-word picture book in five chapters comes to about 40 credits. A 70,000-word novel is roughly 400,000 characters, so about 1,600 credits: about $160 at the 10¢ a credit of the smallest pack, and less on bigger packs.
ElevenLabs’ Creator plan is $22 a month for 121,000 credits at about 1 credit per character, or about 18¢ per 1,000 characters, and the retailers’ tools are free to produce with. Anthra is the more expensive way to voice a full-length novel. If price decides it, use one of them. See Anthra pricing.
No. ACX requires human narration unless ACX has authorized otherwise and prohibits unauthorized AI audio, whatever the file quality. Amazon’s own route for AI narration is KDP Virtual Voice, which uses Amazon’s voices.
No. At 4 credits per 1,000 characters, a 70,000-word novel is about 1,600 credits, or about $160 at the smallest pack’s rate. ElevenLabs direct costs less per character, and KDP Virtual Voice and Google Play’s auto-narration are free to produce with. Anthra’s narration suits picture books and shorter works better.
Only if the book already lives in Anthra and you value the workflow: narration from the same manuscript, a voice per character, a pronunciation list and re-narration of only what changed. For audio alone, ElevenLabs has more voices and costs less per character.
As of November 2026, Google Play accepts third-party AI audio marked “Synthesized voice”. Spotify names providers such as Google Play Books and ElevenLabs. INaudio takes only unmodified LPF files from three named sources. ACX prohibits unauthorized AI audio. Rules change, so read each retailer’s own page first.
A whole-book M4B with chapter markers, and an MP3 for each chapter. The retail audio package adds per-chapter MP3s at 192 kbps CBR with a measured loudness report. We don’t claim any of them meets a retailer’s spec, so compare them with the retailer’s requirements before you upload.
Checked November 2026. Product names are trademarks of their owners; this page is not affiliated with or endorsed by them.
Start with welcome credits, no card required. See pricing, or read about the agentic book creator.
Start your book