KO Reader

KO Reader turns web articles, ebooks, and PDFs into natural-sounding speech — entirely on your device. No cloud, no tracking, no subscriptions. Paste a URL or open an EPUB or PDF, and KO Reader reads it aloud with a neural AI voice. Everything runs locally, so your reading stays private and works anywhere — on the train, on a plane, or off-grid. Export your favorite articles or books to m4a or m4b files. WHAT YOU CAN READ • Web articles — Wikipedia, news sites, blogs, and Substack • Fanfiction — Archive of Our Own, Royal Road, and Wattpad, with chapter detection • EPUB ebooks with full chapter navigation • PDF documents • Plain text — paste or type any content • Safari Share Extension — send articles to KO Reader with one tap NATURAL AI VOICES • 53 neural voices across 9 languages: American and British English, Spanish, French, Italian, Portuguese, Japanese, Mandarin, and Hindi • Preview any voice before selecting it • Mark favorites for quick access • Playback speed from 0.5x to 2.0x • Smart heteronym resolution — the voice knows the difference between "lead" (the metal) and "lead" (to guide), "wound" (injury) and "wound" (past of wind) • Custom pronunciation dictionary — teach the reader how to say names, jargon, or anything you like READ YOUR WAY • Reading queue — line up multiple articles and auto-advance • Bookmarks — save passages to come back to • Reading history — pick up where you left off, across books and articles • Chapter pre-generation — book chapters start instantly, no waiting between them • Audio export — save generated audio for offline listening • Adjustable article font (system, serif, monospaced, rounded) and Dynamic Type support BUILT FOR ACCESSIBILITY • Full VoiceOver support with labeled controls throughout • Haptic feedback on transport controls and sleep timer • Background audio — listen while you use other apps or with the screen off • Sleep timer — stops playback when you drift off • Works with AirPods, Bluetooth speakers, and CarPlay RUNS EVERYWHERE • Universal app for iPhone, iPad, and Mac • macOS keyboard shortcuts (Space to play/pause, Cmd+← and Cmd+→ to skip) • Optimized for Apple Silicon with optional CoreML acceleration • Adjustable CPU thread count for performance tuning PRIVACY FIRST • 100% on-device TTS — your content never leaves your phone • No accounts, no analytics, no tracking, no ads • No internet required after the one-time model download • Reading progress stored locally only KO Reader downloads a ~350 MB neural voice model on first launch. After that, everything works fully offline. Powered by the open-source Kokoro TTS model (Apache 2.0).