Accessibility and reading support
Listen to long-form text when visual reading is difficult, tiring, or not the preferred way to access a document.
Tools / Ebook & Document Audio
Updated August 2026
Turn EPUB, PDF, and TXT reading material into a private read-along audiobook in three simple steps. Parse and store the book in your local bookshelf, generate paragraph audio, and resume in this browser. Non-English Kokoro has the text phonemization exception explained below.
Free · No signup · Works in your browser
Your bookshelf, reading progress, and completed paragraph audio stay in this browser until you remove them.
TXT, EPUB, and text-based PDF support
Private browser parsing and voice generation
Paragraph audio cached as each section finishes
Resume later or continue into a read-along view
EPUB, PDF, TXT, and general document-to-audio searches all lead here because the same importer, local bookshelf, and paragraph generation workflow handles the supported input.
Interactive workspace
Runs in your browser
Follow three simple steps: import a book, choose how it should sound, then generate paragraph audio. Your local bookshelf keeps completed work ready to resume or open in the read-along reader.
Sponsored
Ads help keep OfflineTTS free to use.
Quick answer
An ebook-to-audio converter turns written book or document text into spoken audio. OfflineTTS reads EPUB, TXT, and text-based PDF files in the browser, separates the content into chapters and paragraphs, and generates speech that can be listened to, read along with, resumed later, or exported.
Converting a book to speech is useful when listening is more practical than looking at a screen, or when hearing the text provides a second way to review it.
Listen to long-form text when visual reading is difficult, tiring, or not the preferred way to access a document.
Pair written paragraphs with spoken audio for review, pronunciation practice, and repeated listening.
Hear a manuscript, report, or draft aloud to notice awkward phrasing, repeated words, and missing transitions.
Prepare chapters for listening during travel, walks, chores, or other times when holding a screen is inconvenient.
Turn text-based PDFs and notes into a listening queue while keeping the original wording available in the reader.
Create a private narration draft before recording a final human performance or editing a publication-ready audiobook.
The browser workflow keeps setup short while still allowing chapter selection and optional model controls.
Choose an EPUB, TXT, or text-based PDF. The file is parsed locally and added to this browser's bookshelf.
Match the voice language to the book and select only the chapters needed for the current listening session.
Audio is generated paragraph by paragraph and cached after each result, so completed work is not repeated.
Open the reading view to follow highlighted paragraphs, or download ordered WAV or MP3 paragraph files.
Use the source that contains the cleanest selectable text and chapter structure.
| Format | Best for | How it is handled | Important limitation |
|---|---|---|---|
| EPUB | Novels and structured ebooks | Reads the book spine, metadata, and chapter text | Protected or DRM-locked files are not supported |
| Reports, study packets, and papers | Extracts selectable text page by page | Scanned image PDFs require OCR first | |
| TXT | Plain manuscripts, notes, and drafts | Uses blank lines as paragraph boundaries | Chapter titles may need to be added manually |
Long chapters are divided into natural paragraphs and model-safe text chunks. Each paragraph is saved as soon as its audio finishes. If generation is paused, the tab closes, or one paragraph fails, the local bookshelf can continue with pending and failed paragraphs instead of starting the book again.
Book text, settings, reading position, and generated WAV data are stored in this browser's IndexedDB. They are not a cloud backup and do not automatically appear on another device or browser profile. Clearing site data or deleting a bookshelf item removes the local copy.
Importing an EPUB, TXT, or text-based PDF parses the selected file in the browser and writes the book, chapter structure, reading position, generation state, settings, and completed paragraph audio to local browser storage. OfflineTTS does not create a cloud bookshelf or sync the imported file to another device. A remote backup is therefore not available if you clear site data, delete the book, lose the browser profile, or use a different device.
Speech generation follows the selected engine's data path. English Kokoro, Kitten, Piper, Supertonic, and
Whisper-based local features perform their model work on the device after required assets are available.
Non-English Kokoro sends the paragraph text being generated to
api.offlinetts.com for phonemization, receives a
pronunciation representation, and synthesizes the audio locally. Do not choose that workflow for text that
must remain completely disconnected from a network.
Model downloads, site hosting, and ordinary website analytics are separate network requests described in the Privacy Policy. To remove the local bookshelf, delete individual books in the tool or clear all site storage for OfflineTTS in browser settings. WAV, MP3, or ZIP exports are files outside browser storage and must be deleted separately. Use only books and documents you are authorized to process, and review model and voice terms before distributing generated narration.
Generated books open in a dedicated local reader. It shows one chapter at a time, highlights the active paragraph, skips paragraphs without audio, and saves the current paragraph and playback time for the next visit.
Answers about supported files, long books, local storage, voices, exports, and the read-along reader.
OfflineTTS supports TXT, EPUB, and text-based PDF files in the browser-based ebook workflow.
The selected file is parsed into the local bookshelf in this browser rather than uploaded as a book. Speech synthesis is local; if you choose a non-English Kokoro voice, the paragraph text needed for generation is sent to the documented phonemization endpoint before local synthesis.
Each completed paragraph is cached as WAV audio on this device. Download paragraph files as WAV or MP3, or continue into the read-along experience.
Yes. Your book text, paragraph status, voice settings, and completed audio stay in the local bookshelf in this browser until you delete them.
Not yet. PDF support requires selectable text. Image-only and scanned PDFs need OCR before they can be converted to audio.
Available languages and voices depend on the speech model selected in Advanced settings. Choose the book language first, then select a compatible voice.
Generation time depends on book length, the selected model, and your device. The first run may also download a voice model. Paragraph caching lets you pause and continue without restarting completed work.
Yes. After generating audio, open the local read-along reader to view the current chapter, highlight the active paragraph, and continue from your saved position.