VoiceWeaver

Listen to anything: books, PDFs, articles, and pasted text, read aloud with natural on-device AI voices. Private by design. Your library never leaves your device.

VoiceWeaver turns your documents into audiobooks that are read aloud with natural neural voices that run entirely on your iPhone or iPad.

IMPORT ANYTHING • EPUB books, PDFs (with built-in OCR for scanned pages), Word documents, plain text, and Markdown • Scan paper pages with the camera • Web articles — paste a link and VoiceWeaver extracts the readable text • Share a page or selected text to VoiceWeaver from Safari or any app • Browse and download thousands of free public-domain classics from Project Gutenberg and Standard Ebooks, right inside the app

NATURAL VOICES, YOUR CHOICE • VoiceWeaver TTS: high-quality neural speech generated on-device using the Apple Neural Engine — no internet needed, nothing leaves your phone (Based on the Kokoro TTS model) • Your own Personal Voice • Built-in iOS voices, including Premium and Enhanced • Optional cloud voices: ElevenLabs, OpenAI, and Google Cloud with your own API key, or your self-hosted Chatterbox / Kokoro server

MADE FOR LISTENING • Adjustable speed, skip controls, sleep timer, and bookmarks • Follow along on screen while it reads — the current passage stays highlighted • Background playback with lock-screen and Control Center controls • Progress, time-remaining estimates, and full-text search in every book • Pronunciation overrides: teach it names and unusual words, either with a simple respelling or exact phonetics with the built-in IPA keyboard • Fix source-text mistakes without leaving the app

BUILT AS AN ACCESSIBILITY TOOL • Full VoiceOver support, Dynamic Type, and Reduce Motion throughout • Designed for readers with low vision, dyslexia, or anyone who prefers to listen

PRIVATE BY DESIGN • Your books and reading data stay on your device (with optional iCloud sync to your own devices) • No accounts, no ads, no tracking, no analytics • Text is only ever sent to a cloud voice service if you deliberately configure one; the on-device voice never sends anything anywhere