for PDFs and EPUBs

read and listen to your documents.

PageVoice is an open-source reading app. Import a PDF or EPUB, read it, and listen with a voice that runs on your device. Your files are processed in your browser, and no account is needed.

Open beta Automated tests pass in Chromium. Firefox passes most of them. Safari support is partial.

The PageVoice reader with The Happy Prince by Oscar Wilde open. The contents list is on the left, the chapter text is in the middle with the sentence being read highlighted, and the audio player is on the right.

What PageVoice does

reading

Adjust the theme, text size, and line width to suit you.

  • Three themes, and adjustable text size and line width
  • Contents list, section navigation and bookmarks
  • Find in book ignores accents and highlights every match
  • Focus mode hides everything except the text
  • Remembers your place in each document

listening

Voices run on your device. They are fetched once, the first time you listen, and then work offline.

  • Kokoro and Piper voices, female and male, in English and Spanish
  • The sentence being spoken is highlighted and kept in view
  • Speed from 0.5× to 2×, a sleep timer, and media keys
  • Start listening from any sentence. Audio is prepared ahead.

search and structure

Find text in a book, and see how it is divided into sections.

  • Keyword search across the whole book, with section citations
  • Detected sections and front matter you can review and correct
  • Results are passages from the book. PageVoice does not generate summaries or answers.

Planned, not yet built. AI-assisted analysis built on Claude: source-grounded questions, explanations, and better text preparation for narration. Search today finds passages. It does not answer questions. See the roadmap.

accessibility is the point

Read with your eyes, your ears, or both.

  • Every control works from the keyboard, with visible focus
  • Keyboard shortcuts for play, find, sentence navigation, and focus mode
  • Labeled controls and live status messages for screen readers
  • Text up to 32 px. Three themes. Reduced motion is respected.
  • The spoken sentence is marked by a bar and an icon, not by color alone

local processing, stated plainly

Where documents are processed
Parsing, OCR, speech, and search run in Web Workers inside your browser. PageVoice has no accounts and no server that receives your files.
Where they are stored
In this browser's storage on this device. Clearing site data removes them; library backups are a ZIP file you keep.
What uses the network
Loading the app, fetching the voice from Hugging Face the first time you listen, and fetching text recognition data from jsDelivr the first time you import a scanned PDF. These are fixed files checked against pinned hashes; the requests contain no document text. The app blocks connections to any other server. No analytics.
Narration
Speech is generated in your browser. No sentence of your books is sent anywhere, and the audio is kept in this browser only.

supported formats

Formats PageVoice can open
Format What you get
EPUB 2 and 3 Chapters, title, author and cover. DRM-protected books cannot be opened.
PDF with text Sections come from bookmarks and headings. Repeated headers, footers, and page numbers are removed.
Scanned PDF Readable. English + Spanish text recognition (25 MB) is fetched automatically the first time.

Files up to 100 MB. PageVoice detects each document's language automatically. Narration is available in English and Spanish. The browser needs WebAssembly and IndexedDB.

getting started

  1. open PageVoice

    It runs in the browser. You can also install it as an app.

  2. add a PDF or EPUB

    Choose a file or drop it on the page. You can read right away.

  3. listen

    Press play. The voice (about 120 MB for English, 100–140 MB for Spanish) is fetched once; after that, listening works offline.

Open PageVoice