How to Create an Audiobook with AI

What You'll Need

Start by navigating to the Audiobook Narration page and clicking “Create Your Audiobook.” You'll see a drag-and-drop upload area that accepts five book formats:

Once your book is uploaded and parsed, you'll land in the audiobook workspace. This is your command center for the entire production process. On the left, you'll see a chapter sidebar listing every chapter detected from your book:

Each chapter shows:

• Chapter title (editable — click to rename)
• Word count for estimating narration length
• Status indicator: green (audio generated), orange (content changed since last generation), gray (not yet generated)

You can reorder, duplicate, delete, and create new chapters directly from the sidebar. The main content area shows a rich text editor for the selected chapter — powered by Tiptap — where you can edit your text, format with bold/italic/underline, and add bullet lists.

At the top of the workspace, you'll find the toolbar with the main voice selector, generate button, download options, and the characters panel toggle:

Audiobook Workspace Toolbar

Aoede

en-US · Female

Characters5

248,350

GenerateDownload

The toolbar gives you quick access to:

• Main voice selector — sets the default narrator voice for the entire book. Click to change it. This voice is used for all narration paragraphs unless overridden by character detection.
• Characters button — opens the character detection panel (covered in Step 3). The badge shows how many characters have been detected.
• Credits balance — your remaining generation credits.
• Generate button — click to generate audio. A dropdown lets you choose between generating the current chapter or the full book.
• Download button — export options including per-chapter MP3, merged full book, and ZIP archive.

This is where the magic happens. Click the “Detect Characters” button, and Notevibes AI scans your entire manuscript to identify every speaking character. The AI doesn't just find names — it understands narrative roles, counts dialogue lines, maps every paragraph to the right speaker, and automatically detects your book's genre to shape the narration style. No manual preset selection needed — the AI figures out whether your book is a thriller, romance, fantasy, or memoir and adapts the delivery accordingly.

Characters — The Midnight Garden

Detected Characters5

Detect Characters

Narratornarrator

Voice: Aoede142 lines

Elenaprotagonist

Voice: Kore87 lines

Marcusantagonist

Voice: Orus54 lines

Old Keepermentor

Voice: Puck31 lines

Rinnsidekick

Voice: Leda23 lines

Apply to All Chapters

For each detected character, the AI determines:

• Name and any aliases used in the text
• Role classification: narrator, protagonist, antagonist, mentor, sidekick, supporting, or minor
• Gender (for voice matching)
• Dialogue line count and chapter appearances
• Narrative arc and character evolution across the book
• A suggested AI voice based on the character’s attributes
• Per-paragraph delivery directions (“whispered, trembling”, “cold, controlled”) generated from the scene context

The characters panel is fully editable. Rename characters, change their roles, swap voices, or remove false detections. When you're happy with the lineup, click “Apply to All Chapters” to propagate the voice assignments across your entire book.

The AI also auto-detects your book's genre (fiction, thriller, romance, fantasy, sci-fi, memoir, etc.) and uses it to shape character portraits, scene illustrations, and the overall visual theme if you later publish with Storybook Mode. There's no manual genre or narration style selector to worry about — the AI reads your text and adapts automatically.

Each character needs a distinct voice. The AI suggests voices automatically based on gender, role, and personality, but you can override any assignment. Click on a character's voice to open the voice selector:

Select Voice

AllFemaleMale

AllStarredRecentNatural

Aoede

Female · en-US

Kore

Female · en-US

Leda

Female · en-US

Orus

Male · en-US

Puck

Male · en-US

Charon

Male · en-US

550+ voices across 57 languages

The voice selector offers 550+ AI voices organized by:

• Gender: Male, Female, or All
• Type: All, Starred (your favorites), Recent (last 5 used), Natural (Chirp3-HD), Standard
• Language: 57 languages including regional accents (American, British, Australian, Indian, Scottish, Nigerian, and more)
• Preview: Click the play button on any voice card to hear a sample before selecting

The latest Chirp3-HD voices are the most natural-sounding, with 30 unique voice profiles named after astronomical objects: Aoede, Kore, Orus, Puck, Leda, Charon, and more. Each one supports all 57 languages natively — meaning the same voice can narrate your book in English, French, Japanese, or Arabic with authentic accents.

After character detection, every paragraph in your book gets assigned to a specific voice. The paragraph gutter — the left-side overlay in the editor — gives you a visual map of who's speaking where:

Chapter 3 \u2014 The Letter

Narrator

The rain hammered against the window. She knew, even before opening the letter, that everything was about to change.

Narrator

Reflective, slow — building tension

The envelope was thin — just a single sheet folded in thirds. No return address. Her name in handwriting she hadn’t seen in seventeen years.

Elena

Whispered, trembling

“Wait,” she whispered, pressing the paper flat on the table. “This can’t be right. Unless... unless he knew all along.”

Marcus

Cold, controlled

“You weren’t supposed to find that.” His voice came from the doorway, flat and measured. “Not yet.”

The paragraph gutter shows:

• Voice avatar: A visual indicator of which character is speaking this paragraph. Click to change the voice.
• Speaker name: The character name displayed below the avatar.
• Sync status: Green dot = audio matches current text. Orange dot = text has changed since audio was generated.
• Play button: Click to hear this specific paragraph. The currently playing paragraph is highlighted.
• Scene direction: When Adaptive Delivery is enabled, each paragraph shows an italicized delivery cue (e.g., “Whispered, trembling” or “Cold, controlled”).

Click the “Generate” button in the toolbar to start producing your audiobook. A dropdown gives you two options:

When Character Voices is enabled, clicking generate will first run AI character detection (if characters haven't been detected yet), then generate audio with per-character voice assignments. When Adaptive Delivery is on, the AI computes scene-aware delivery directions for each paragraph before synthesis.

After clicking generate, the progress dialog opens and shows real-time status for each chapter:

Behind the scenes, Notevibes processes each chapter through a sophisticated pipeline: paragraphs are grouped into voice segments, chunked for the TTS engine, and synthesized with per-character voices and delivery directions. For dialog-heavy scenes, the engine batches consecutive short dialogue segments from two speakers into a single multi-speaker request — producing more natural conversational flow with fewer audio stitching artifacts.

After generation completes, every paragraph becomes playable. Click the play button in the paragraph gutter to hear individual paragraphs, or use the audio player bar at the bottom to play through the entire chapter.

The editing workflow is non-destructive:

• Edit text freely — the orange sync indicator tells you which paragraphs need re-generation
• Change a character’s voice — only the paragraphs using that voice are marked stale
• Adjust delivery directions — toggle Adaptive Delivery on/off or edit individual directions
• Regenerate selectively — you can regenerate a single chapter without touching the rest
• The generated audio for unchanged chapters is preserved and cached

Once your audio is generated, you can bring your audiobook to life visually. Notevibes generates two types of AI illustrations: character portraits and scene illustrations. These power Storybook Mode when you publish your book.

Character Portraits

Character Portraits3 of 5 ready

Generate All

Narrator

Aoede

Elena

Kore

Marcus

Orus

Old Keeper

Puck

Rinn

Leda

Character portraits are generated using your book's character descriptions, genre, and role classifications. The AI applies genre-specific visual styles:

Scene illustrations are generated at key narrative moments in each chapter. The AI uses the paragraph text, delivery directions, and character information to create contextual illustrations. These appear on the left page in Storybook Mode's Illustrated layout, creating a book-spread experience where listeners see the scene while hearing the narration.

Once you're happy with the narration, click the download button. Three export options are available:

Every export is broadcast-quality. Notevibes handles volume normalization, peak limiting, and format encoding automatically — no audio engineering tools or post-processing needed. The files are ready to upload directly to Audible, Google Play Books, Apple Books, or any other audiobook distributor.

You have two publishing paths: publish directly on the Notevibes Library for instant availability with rich features, or export and distribute to external platforms.

Publish on Notevibes Library

When you generate a book, it's automatically added to your library as a private draft. To publish publicly:

1. 1.Open the book in your Library and fill in metadata: author, description, genre, language, tags
2. 2.Optionally generate AI character portraits and scene illustrations
3. 3.Enable Read Along (word-level sync) for interactive reading
4. 4.Click “Publish” to make the book available on the Notevibes Library
5. 5.Share the public URL with your audience

Published books on Notevibes come with Storybook Mode, Read Along, character portraits, scene illustrations, and listener analytics included — features you won't find on any other audiobook platform.

Export to External Platforms

Download your audio files and distribute to any major platform:

What Your Listeners Experience

Publishing on the Notevibes Library gives your audience three ways to experience your audiobook — features you won't find on any other platform. These are powered by the character portraits, scene illustrations, and audio sync you generated in the previous steps.

Storybook Mode

Transforms your audiobook into a visual reading experience. Listeners see beautifully typeset pages with the narration synced to the text — a hybrid of audiobook and illustrated storybook.

The Midnight Garden

Chapter 7 \u2014 The Crossing

The Midnight GardenChapter 7

Elenaprotagonist

The forest was silent save for the distant call of a nightingale. Elena pressed her palm against the ancient oak, feeling the bark rough beneath her fingers.

“We shouldn't have come here,” Marcus whispered from behind the thornwall. “Not tonight.”

She didn't answer. Instead, she stepped forward, through the archway of twisted branches, into the moonlit clearing.

— 127 —

4:12

11:05

Read Along

Word-level highlight sync that follows the narration in real time. Each word lights up as it's spoken, making it easy to follow along.

Listener Analytics

Track how listeners engage with every chapter of your audiobook with the built-in analytics dashboard.

Plus geographic insights (listener locations by country), device & behavior data (mobile vs. desktop, session length, chapter drop-off points), and daily aggregation for trend analysis.

Traditional Studio vs. AI Narration

A 10-hour audiobook produced in a traditional studio would cost between $50,000 and $150,000, take weeks of studio time, and be locked to a single language. With Notevibes, the same audiobook can be created in an afternoon, revised unlimited times, and published in 57 languages — all from a single manuscript.

Frequently Asked Questions