Chapter breaks
Chapters are detected and named for you, so the finished file has a marker at every chapter instead of one long track.
Audiobook maker
Upload a manuscript or paste a story, and this audiobook maker does the studio work: it finds the chapters, gives the narrator and each speaking character an AI voice of their own, and exports a chaptered M4B or one MP3 per chapter. You direct the lines that matter and can fix any single line without touching the rest.
EPUB, PDF, DOCX, TXT or MD · up to 30 MB
Read in your browser. Nothing is uploaded until you generate
No microphone, no booth and no audio editor. Most of your time goes into listening, not producing.
Drop in an EPUB, Word file, text-based PDF, TXT or Markdown file up to 30 MB, or paste a chapter. The audiobook maker opens the file in your browser first.
Chapters are found from the contents or your headings, and the cover, copyright page and dedication are set aside before narration starts.
Speaking characters are picked up from dialogue tags such as “said Mara”. Keep the suggested voices or swap any of them, then add a [direction] where a scene needs it.
Generate one chapter to test, then the rest. Download an M4B with chapter markers, an MP3 per chapter or a ZIP of every file.
Every format ends up as the same chapter list inside the audiobook maker, so casting and export work the same way whatever you start from.
| You upload | How chapters are found | Worth knowing |
|---|---|---|
| EPUB 2 or EPUB 3 | The table of contents (nav or NCX) | The cleanest start; the file must be DRM-free |
| Word (DOCX) | Heading 1 and Heading 2 styles | Bold text that only looks like a heading is not a break |
| Text-based PDF | Reflowed text, chapter headings kept | Page numbers and running headers are dropped; scans are not read |
| TXT or Markdown | Lines like “Chapter 3”, “Part Two” or # headings | Plain exports from Scrivener, Obsidian or any editor work |
| Pasted text | The same rules as a TXT file | Handy for a short story, a scene or a single chapter |
Format notes in depth: EPUB to audiobook and PDF to audiobook.
Reading text aloud is the easy part. A good audio book maker also handles the jobs that eat a weekend when you stitch tools together yourself.
Chapters are detected and named for you, so the finished file has a marker at every chapter instead of one long track.
Copyright notices, contents lists, “also by” pages and dedications are not narrated, so listeners start on page one of the story.
A narrator reads everything outside quotation marks, and up to ten named characters get their own voices, plus an “Other voices” role for minor speakers.
A short note in brackets, such as [whisper] or [furious], changes how a line is performed, without any audio editing.
A wrong name or a flat reading means regenerating that one line. The rest of the chapter keeps the audio it already has.
Files are parsed on your device, only the lines you generate reach the speech model, and manuscripts are never used to train AI models.

Same cast in every bookSingle-voice text to speech reads the hero and the villain in the same tone. This audiobook maker follows the dialogue tags in your text, works out who is talking and gives each speaking character a separate voice, matched on gender where the text makes it clear. The narrator voice carries everything else.
You stay in charge of the casting. Pick from 137 AI voices for audiobooks, lock the cast and reuse it for the next book in a series, so recurring characters sound the same in book four as in book one. If a line is given to the wrong speaker, reassign it in the studio and regenerate only that line.
Pick a genre and play the voices it tends to cast for narrators, heroes and villains. Every role can be recast before you generate.




A narrator built for long quests, a raider who sounds the part and a sorceress with secrets.
Premium voicesTap a role to hear its voice.
Fantasy audiobook narration guideAn audiobook maker like this suits most fiction and a lot of nonfiction. A few projects still need a different route.
What it costs. Importing a book, detecting chapters and casting are free. Audiobook makers price in different ways; here audio is paid in credits per 1,000 characters, about $9 to $12 per finished hour with Standard voices, so a 9 hour novel lands near $80 to $110. Sign up with Google and you get 100 gift credits (valid for 7 days, about 12 minutes on Standard; usage varies by text and voice tier) for a short sample. Put your own word count into the audiobook cost calculator or compare plans on the pricing page.
Which voice model to choose. Standard voices suit most books. Lite is cheaper for drafts and long nonfiction, Premium gives you 67 ElevenLabs character voices, and Expressive adds laughs, whispers and the widest emotional range at a higher credit rate. You can switch models per project, so a free audiobook maker trial on Standard tells you a lot before you pay for anything.
Which file to export. An M4B is one file with a marker at every chapter, which audiobook apps love; the M4B audiobook maker page explains what goes into it. MP3 files suit stores, podcast hosts and older players. You can export both from the same audio.
Where you can publish. You own the audio you generate from text you hold the rights to. Store rules for AI narration differ and change, so read where to sell AI audiobooks and check each store’s current policy before you upload. Selling direct from your own site is always open.
Books in other languages. The voice library covers 13 languages, including Spanish, German, French, Japanese and Hindi, with English the most polished. Cast voices from the language of your text, and keep one language per project for the most natural result.
Uploading, chapter detection and casting are always free, and new Google sign-ups get free credits to generate a short sample with no card needed. That is enough to hear your narrator and main characters before deciding whether to make the whole book.
No. The audiobook maker generates every line with AI voices, so there is nothing to record, clean up or master. Your job is to check the chapters, approve the cast, add a few directions and listen through before you export.
Yes. Every book you start in the AI audiobook generator is saved in “My Audiobooks” in your browser with its chapters, cast and finished audio. Come back days later, open the project and generate the next chapters with the same voices.
Yes. Click a line in the studio, change the words or add a direction, and regenerate only that line. The rest of the chapter keeps its audio, so late fixes cost a few credits rather than a whole chapter.
It runs in any modern browser, including Safari on iPhone and Chrome on Android. Projects are stored in the browser you used, so a book started on your laptop stays on that laptop unless you export the files.
Short stories, memoir, how-to books, bedtime stories and serial fiction all work. People also voice AI bedtime stories for children and table reads of their own drafts. Anything with clear text and chapter headings can become an audiobook.

Upload your book, hear the full cast in minutes and generate a short sample free, no card needed.