Can You Sell an AI-Narrated Audiobook of a Public-Domain Book?

Yes — if a book's text is in the public domain where you publish, you can legally produce and sell an audiobook of it, including an AI-narrated one, with no author permission. You must verify public-domain status for your jurisdiction, use a clean public-domain edition, and disclose AI narration wherever the retailer requires it.

Public-domain titles are the foundation of a low-risk audiobook catalog: classics like Dracula, Pride and Prejudice, The Great Gatsby, and the early Sherlock Holmes stories have no living rightsholder to license, so anyone can record and sell them. AI narration makes the economics work, because producing audio of a 100,000-word novel for a flat fee instead of thousands of dollars is what turns a single public-domain title into a viable catalog play. This guide explains exactly when that is legal, how to verify it, the one copyright nuance that catches catalog builders out, and where you can actually sell the result.

This post is about publishing books you did not write — works whose copyright has expired. If you wrote the book yourself and want to know whether you can commercially sell an AI narration of your own title, that is a different question answered in the AI audiobook commercial rights guide. Here the text belongs to no one, which removes the licensing question but introduces a verification question and a competition question instead.

Disclaimer: This article is general information, not legal advice. Copyright terms differ by country and edition, and the rules around AI authorship are still developing. Consult a qualified intellectual-property lawyer before building a catalog on any specific title.

Is it legal to sell an AI audiobook of a public-domain book?

Yes, selling an AI-narrated audiobook of a public-domain book is legal, because public domain means copyright in the text has expired and no permission or licence is required to reproduce, adapt, perform, or sell it. Recording an audiobook is a form of adaptation, and once a work is in the public domain, that right is open to everyone. The catch is not whether you may do it — you may — but whether the specific edition you use is genuinely public domain, and whether your narration method satisfies the retailer's disclosure rules.

The most common mistake is producing audio from a copyrighted version of a public-domain work. The original 1851 text of Moby-Dick is public domain, but a modern annotated edition, a fresh introduction, or a recent translation of a foreign-language classic each carries its own new copyright. If you narrate a translator's words, you are using a protected work even though the underlying story is centuries old. Always source the audiobook from a clean public-domain edition — the original text or a translation old enough to be out of copyright itself.

How do I verify a book is in the public domain?

Verifying public-domain status comes down to jurisdiction and dates, and the rule differs between the United States and most of the rest of the world. You verify by identifying the publication date, the death dates of every contributor, and the law of the country you are publishing in — then confirming the result against an authoritative public-domain reference before you produce anything.

In the United States, the rule is date-of-publication based: as of 2026, every work published in 1929 or earlier is in the public domain, and the cutoff advances by one calendar year every January 1 (works from 1930 entered the public domain on Public Domain Day, January 1, 2026). Because the line moves annually, always re-check the current year's cutoff rather than relying on a number you memorised. The Duke University Center for the Study of the Public Domain publishes the current cutoff each Public Domain Day, and the U.S. Copyright Office duration FAQ explains the underlying terms.

In the European Union, the United Kingdom, and most of the world, the rule is life-of-the-author based: copyright lasts for the author's lifetime plus 70 years. A book is public domain in the EU only once 70 full years have passed since the author died, which means a work can be public domain in the US but still protected in Europe, or vice versa. Because a translator, illustrator, or editor is a separate author with their own term, you must confirm the death date of every contributor whose work appears in your edition, not just the original novelist.

The takeaway from the checklist: a title is only safe to produce when the text, the specific edition, and every contributor all clear the public-domain bar in your selling market. If any one of them is still in copyright, you are working from a protected source even if the story itself is famous and old.

Does anyone own the copyright in an AI-narrated public-domain audiobook?

The copyright status of the AI narration is the nuance catalog builders most often miss. When a human narrates a public-domain book, the resulting sound recording is a new copyrightable work — the performance is the human authorship that earns protection, even though the text underneath is free. But the U.S. Copyright Office has stated that material generated purely by AI, without meaningful human authorship, is not protected by copyright. So a purely AI-generated narration of a public-domain text may have no copyright protection of its own.

The practical consequence is competition. Because the text is public domain and a purely AI-generated recording may not be protectable, another publisher can legally produce their own AI audiobook of the exact same title — there is no exclusive recording to infringe. You cannot fence off "the AI audiobook of Frankenstein." What you compete on instead is voice quality, production polish, accurate pronunciation of period and proper nouns, cover and metadata, and discoverability. The Copyright Office's position on AI and human authorship is set out in its Copyright and Artificial Intelligence guidance, and the registration policy is summarised in the Office's 2023 registration guidance for AI-generated material.

A non-copyrightable narration does not stop you selling the audiobook or earning from it — sales revenue does not depend on holding a copyright in the recording. It simply means your edge is the product itself, not legal exclusivity. For most catalog builders that is fine: the goal is a steady-selling, well-produced edition of a beloved classic, not a monopoly on it.

Where can I sell a public-domain AI audiobook?

A public-domain AI audiobook is an author-supplied AI audio file, so it reaches retail the same way any external AI audiobook does: directly to the two most open platforms, and to the rest through an aggregator that accepts AI narration. The table below maps each route, and it carries one extra caveat that applies specifically to public-domain titles — some stores limit how many editions of the same classic they will list.

The takeaway from the table: you can upload directly to Google Play Books and Kobo, and reach Apple Books, Spotify, Chirp, and libraries through an aggregator such as PublishDrive or Author's Republic that accepts AI narration. Standard ACX requires human narration and is not open self-service for external AI files, so a public-domain AI audiobook does not go through the standard ACX portal — the rules are laid out in the ACX requirements guide. For the full distribution map of any AI audiobook, see where to sell an AI audiobook.

Because public-domain classics attract many competing editions, selling direct from your own store is especially attractive for this catalog: there is no duplicate-listing limit, you keep the largest share of revenue, and you own the customer relationship. The mechanics of that are in how to sell audiobooks direct.

Do I still have to disclose AI narration on a public-domain title?

Yes. The disclosure requirement attaches to the narration method, not to the authorship of the text, so a public-domain book narrated by AI must be labelled as digital or synthesized voice exactly as an original title would be. Public-domain status changes who owns the words; it does not change how retailers treat AI-generated audio.

In practice that means Kobo asks you to set the narrator field to "Synthesized Voice," aggregators carry an AI-narration flag in their metadata that feeds Apple and Spotify, and ACX and Audible require AI disclosure in the book description. Treat disclosure as non-negotiable on every platform: it is a platform rule that applies equally to public-domain and original audiobooks, and getting it wrong is a faster route to a takedown than any copyright issue.

How do I produce a public-domain audiobook with AI?

Once you have verified the title is public domain, producing the audiobook is a short, repeatable workflow. The flat-fee, length-independent economics are what make a public-domain catalog viable, since classics are often long and per-finished-hour human narration would cost thousands per title.

1. Confirm public-domain status for your selling market using the verification checklist above, and clear every contributor in your chosen edition.

2. Get a clean source text from Project Gutenberg or Standard Ebooks, and strip out any modern, still-copyrighted front matter, introductions, or annotations.

3. Choose a voice that fits the era and genre — a measured voice for 19th-century literary fiction, something brisker for a 1920s mystery. The choosing an audiobook voice guide covers how to match voice to material.

4. Run a free first-chapter preview to hear the actual voice on the actual text, and re-generate any chapter at no extra cost if a period name or archaic word is mispronounced.

5. Add accurate metadata and disclose AI narration, then upload to Google Play and Kobo directly and distribute wide through an AI-friendly aggregator.

The whole pipeline, from manuscript to distributable files, is documented in the AI audiobook production guide. The key advantage for catalog builders is that a 150,000-word classic costs the same flat fee as a short novella, so length never blocks a title from being worth producing.

Where TomeVox fits for public-domain catalogs

TomeVox turns a clean public-domain text into a finished audiobook for a flat early-bird fee — $49 up to 60,000 words, $79 up to 100,000 words, and $99 up to 150,000 words, with $0.0005 per word only above 150,000 — and delivers an M4B file with chapter markers plus per-chapter MP3 files, usually within 48 hours. Because the price is flat regardless of length, a long classic costs the same as a short one, which is exactly what makes a multi-title public-domain catalog affordable.

You receive full commercial rights on delivery with no exclusivity, so you can upload directly to Google Play and Kobo and distribute wide through an AI-friendly aggregator, and TomeVox supports 13 languages for foreign-language classics whose text is out of copyright. Every audiobook is automatically checked for technical quality before delivery, a free first-chapter preview lets you hear the voice with no credit card, and you can re-generate any chapter at no extra cost when a period proper noun needs a fix. TomeVox does not provide legal advice or verify public-domain status for you — confirming a title is in the public domain in your market remains your responsibility, and for anything uncertain you should consult an intellectual-property lawyer.

Frequently asked questions

Can I legally sell an AI-narrated audiobook of a public-domain book?

Yes. If the underlying text is in the public domain in the country where you publish, you can produce and sell an audiobook of it, including an AI-narrated one. No author permission or licence is needed because copyright in the text has expired. You must still confirm public-domain status for your jurisdiction, use a clean public-domain edition rather than a copyrighted modern translation or annotated version, and disclose AI narration where the retailer requires it. This is general information, not legal advice.

How do I verify that a book is in the public domain?

Public-domain status depends on jurisdiction. In the United States, works published in 1929 or earlier are in the public domain as of 2026, and the cutoff advances by one year every January 1. In the EU, the UK, and most of the world, copyright lasts for the life of the author plus 70 years, so you confirm the death date of every author, translator, and editor involved. Check a public-domain reference such as the Duke Center for the Study of the Public Domain, the source edition on Project Gutenberg or Standard Ebooks, and the death dates of any contributors before producing the audiobook.

Can someone else narrate the same public-domain book?

Yes. Because the text is public domain, anyone can produce their own audiobook of it. A human-narrated audiobook is a copyrightable sound recording, but the U.S. Copyright Office has stated that purely AI-generated output, with no meaningful human authorship, may not be protected by copyright on its own. That means another publisher could legally produce a competing AI audiobook of the same public-domain title. You compete on voice quality, production, packaging, and discoverability rather than on an exclusive right to the recording.

Where can I sell a public-domain AI audiobook?

An author-supplied AI audiobook file can be uploaded directly to Google Play Books and Kobo Writing Life, and distributed to Apple Books, Spotify, Chirp, and libraries through an aggregator that accepts AI narration, such as PublishDrive or Author's Republic. Standard ACX requires human narration and is not open self-service for third-party AI files. Some retailers limit how many copies of the same public-domain title they list, so check each platform's public-domain policy. Disclose digital-voice narration everywhere.

Do I have to disclose AI narration on a public-domain audiobook?

Yes. Disclosure rules apply to the narration method regardless of who wrote the text, so a public-domain title narrated by AI must still be labelled as digitally or synthesized voice where the retailer requires it. Kobo asks you to label the narrator Synthesized Voice, aggregators carry an AI-narration flag in their metadata, and ACX and Audible require AI disclosure in the description. Disclosing AI narration is a platform requirement, not a public-domain one, and it applies to public-domain and original titles alike.