- Word-by-word highlight — English playback now highlights the exact word being spoken as it reads, not just the current sentence. - Live Activity for audio preparation — preparing a long Chinese text for offline listening now shows progress and chapter position on your Lock Screen and Dynamic Island. - Fixed imported Chinese text files sometimes being misdetected and read in English.

### What's New (1.2.5) - Share text or a web link from any app — Safari, Mail, Notes, and more — straight into your reading library using the system Share Sheet. - Better Chinese novel extraction — short dialogue and sound-effect lines are no longer dropped, and the Safari extension recognizes more chapter page formats.

### What's New (1.2.4) - MOSS is much faster — the Chinese voice engine now runs on a redesigned single-pass model, cutting generation time well below real-time after a brief on-device warm-up. - More consistent voice — tuned sampling keeps tone and pacing steadier from one segment to the next, so your chosen voice sounds more like itself throughout an article.

### What's New (1.2.3) - English is now powered by PocketTTS — faster startup and more natural sound than Otosaku, with voice cloning support. - Voice cloning is now available for English: record a short clip in Settings → Voice Reference and PocketTTS reads in your voice. - Try English voice cloning on the subscription screen — record and hear a sample in your voice before buying anything.

### What's New (1.2.2) - Record a short clip on the subscription screen and hear MOSS read a sample paragraph in your voice — before buying anything. - Opening a Chinese file while Otosaku is selected now shows a one-tap "Use MOSS" prompt so you never get a confusing error. - Choosing Chinese during onboarding now shows the MOSS model download inline, so there are no surprises when you start listening.

### What's New (1.2.1) - MOSS voice models now download on demand — the initial app download is ~30 MB instead of ~650 MB. Download MOSS in Settings → Voice Engine when you're ready. - Short phrases merged for smoother synthesis now each appear on their own line in the reader, so you can follow along phrase by phrase. - MOSS is now the sole on-device Chinese voice engine, with full voice cloning support. Settings are simpler — pick a built-in voice or record your own in Settings → Voice Reference.

### What's New (1.1.10) - Record your own voice directly in Settings → Voice Reference. MOSS encodes it on-device instantly — no Mac or Python setup needed. - MOSS Chinese text-to-speech is significantly faster — audio generation time reduced by up to 35%, so playback starts sooner and offline preparation finishes quicker. - The Safari extension now recognizes more Chinese novel sites — tap the extension icon on chapter pages to send content directly to the app.

### What's New (1.1.9) - MOSS Chinese text-to-speech is significantly faster — audio generation time reduced by up to 35%, so playback starts sooner and offline preparation finishes quicker. - Synthesized audio is now louder and clearer, making it easier to listen in noisy environments without maxing out device volume. - The Safari extension now recognizes more Chinese novel sites — tap the extension icon on chapter pages to send content directly to the app.

### What's New (1.1.8) - 12 fonts now available via a dropdown picker in the reading appearance panel: Chinese options (PingFang SC / 苹方, Songti SC / 宋体, Kaiti SC / 楷体, Yuanti SC / 圆体, Fang Song / 仿宋) and English classics (Arial, Helvetica, Times New Roman, Georgia, Verdana, Courier New). - Special symbols common in Chinese web novels are now sanitized before synthesis so the voice engine never produces noise for unrecognized characters.

### What's New (1.1.7) - Punctuation-only segments (ellipses, dashes, etc.) are now silently skipped instead of producing unexpected sounds during playback. - Very short phrases are automatically merged with the following sentence so the voice engine reads them naturally, without mispronouncing trailing punctuation marks.

### What's New (1.1.6) - Choose text size (small, medium, large) and font style (Default, Serif, Rounded, Mono) from the new "Aa" reading appearance panel in the reader toolbar. - Background color, text size, and font are now all in one place — tap "Aa" while reading to adjust without leaving the page. - New "Browse Chinese Novels" button on the home screen opens a novel site in Safari for easy downloading and importing.

### What's New (1.1.5) - Choose from Paper, Warm, Sepia, Sage, or Night backgrounds in Settings to reduce eye strain during long reading sessions. - Tap the pause button in the reader toolbar to stop background audio preparation at any time. Tap resume to continue. Useful for reading without TTS running in the background.

### What's New (1.1.4) - Imported texts now display as full paragraphs with the active chunk highlighted in yellow, so you can read along as the app speaks. - Chapter navigation works for TXT files — tap the chapter list icon to jump to any chapter, including Chinese web novel formats like "第 1 章". - The Delete Audio button is directly visible on each saved text card — no long-press needed.

### What's New (1.1.3) - The reading highlight now advances instantly on long texts — only the two rows that change redraw instead of all visible rows. - Opening a fully prepared long text is now instant — segment count is read from the manifest instead of scanning every audio file. - MOSS now defaults to the Female 1 voice so you get natural-sounding speech right away without visiting Settings.

### What's New (1.1.2) - Save multiple named voice references for MOSS. Name each one and switch between them in Settings → Voice Reference. - Add a new voice by recording a short clip directly in the app or importing any audio file — MOSS will match your voice style. - The setup flow now lets you choose your voice engine (MOSS, Qwen, or Otosaku) before you start listening.

### What's New (1.1.1) - A new on-device Chinese voice engine powered by MOSS-TTS-Nano CoreML models. Switch to MOSS in Settings for natural Chinese speech with faster startup. - Customize how MOSS sounds by importing a short audio clip or recording your own voice directly in Settings. - Offline audio preparation for MOSS now continues even when you leave the app, so your audio is ready faster.

### What's New (1.0.11) - Resuming prepared audio deep into a long text now loads instantly, no matter how far you've gotten. - Audio segments are now prepared in parallel, so offline preparation finishes faster. - Qwen voices sound cleaner and more natural with tuned generation settings.

### What's New (1.0.10) - Import EPUB, PDF, Markdown, and HTML files directly into your library alongside plain text. - Tap the speed button in the playback bar to adjust reading speed: 0.75×, 1×, 1.25×, 1.5×, or 2×.

### What's New (1.0.9) - Tap the bookmark button while reading to save your place. View all bookmarks and jump to any one instantly. - Tap the chapter list button to see all chapters in a saved text and jump to any one instantly. - Tap the chapter list button to see all chapters in a saved text and jump to any one instantly. - Audio preparation and playback now run in parallel, reducing the pause between paragraphs. - Choose a genre in Settings to match the reading tone to your content — romance, mystery, sci-fi, and 13 more styles. - The voice model now stays loaded across paragraphs, reducing pauses and keeping the tone consistent from start to finish.

### What's New (1.0.6) - Text search now runs in the background with a short delay, keeping the app responsive while you type — even in long documents. - The playback bar now has skip-back and skip-forward buttons to jump between paragraphs without tapping the list. - Saved text cards are now cleaner: delete and audio actions moved to a long-press menu. - AirPods control playback: single-click to play or pause, double-click to skip forward a paragraph, and triple-click to go back. - Playback controls appear on the lock screen and in Control Center.

### What's New (1.0.4) - Fixed an issue where the app could stop responding after Chinese audio preparation was interrupted in the background. The app now detects this automatically and prompts you to restart — your reading position is saved.

### What's New (1.0.3) - Added iCloud sync: imported texts and prepared audio are now shared across your devices. Enable it in Settings → iCloud Sync. - Prepared audio from one device is available to download on another, so you don't need to re-synthesize on every device. - Documents imported on one device appear automatically in the library on your other devices.

### What's New (1.0.2) - Added Chinese text-to-speech using the on-device Qwen model — read Chinese articles and imported text without an internet connection. - Added on-screen guidance for Chinese text files: the app now explains that the on-device model can be slow and recommends preparing audio ahead of time for a smoother experience. - Added a reminder to keep the app in the foreground while audio preparation is running.

## What's New (1.0.1) - Fixed blank Reader and Settings tabs that appeared grey when selected. - Tap any paragraph to jump playback to that point in the article.