Supertonic v3 and Kokoro voices models can now be used on devices with as little as 6GB of RAM like the iPhone 15. Customer voice builder files (voice clones) that you created with Supertone's voice builder tool can now be re-export to work with supertonic v3 and used in Listen2. Also in v1.10: SUPERTONIC V3 The new Supertonic engine fixes the word and phrase skipping that the previous version shipped with. Pacing is rock-solid and latency is lower (12 denoising steps as default now, down from 32). Users with the previous Supertonic engine will see a one-time prompt at launch to download the new model (about 178 MB). 13 SUPERTONIC LANGUAGES Up from 5. Adds Italian, German, Swedish, Greek, Vietnamese, Russian, Turkish, Hindi, and Hungarian. Existing English, Spanish, French, and Portuguese voices carry over. RESUME MID-PARAGRAPH Pausing four minutes into a long PDF page used to restart the whole paragraph on reopen. Listen2 now remembers the sentence you paused on and resumes from there. EXPORT TEXT Long-press a document in the library and tap Export Text to share the extracted text as a .txt file. Useful for sending a chapter to another app, AirDropping a document, or keeping a plain-text backup. Also available as a VoiceOver custom action. BLUETOOTH PAUSE FIX Connecting AirPods mid-playback no longer pauses the audio. Pause now triggers only when the device you were actually listening on disconnects. SHARE EXTENSION Sharing .txt and .md files from Files.app now works directly. Previously these failed with a download error. Unsupported file types now show a clear message directing you to Open in Listen2. SYNC RELIABILITY Books no longer briefly appear as needing re-import after pausing and quickly leaving the library. A race between iCloud sidecar updates and the local document store has been resolved. VOICE LIBRARY CLEANUP The voice library now hides voices whose engine your device cannot run, so you no longer see voices that will not play.

SUPERTONIC V3 The new Supertonic engine fixes the word and phrase skipping that the previous version shipped with. Pacing is rock-solid and latency is lower (12 denoising steps as default now, down from 32). Users with the previous Supertonic engine will see a one-time prompt at launch to download the new model (about 178 MB). 13 SUPERTONIC LANGUAGES Up from 5. Adds Italian, German, Swedish, Greek, Vietnamese, Russian, Turkish, Hindi, and Hungarian. Existing English, Spanish, French, and Portuguese voices carry over. RESUME MID-PARAGRAPH Pausing four minutes into a long PDF page used to restart the whole paragraph on reopen. Listen2 now remembers the sentence you paused on and resumes from there. EXPORT TEXT Long-press a document in the library and tap Export Text to share the extracted text as a .txt file. Useful for sending a chapter to another app, AirDropping a document, or keeping a plain-text backup. Also available as a VoiceOver custom action. BLUETOOTH PAUSE FIX Connecting AirPods mid-playback no longer pauses the audio. Pause now triggers only when the device you were actually listening on disconnects. SHARE EXTENSION Sharing .txt and .md files from Files.app now works directly. Previously these failed with a download error. Unsupported file types now show a clear message directing you to Open in Listen2. SYNC RELIABILITY Books no longer briefly appear as needing re-import after pausing and quickly leaving the library. A race between iCloud sidecar updates and the local document store has been resolved. VOICE LIBRARY CLEANUP The voice library now hides voices whose engine your device cannot run, so you no longer see voices that will not play. CUSTOM SUPERTONIC VOICES Per Supertone, voice embeddings purchased through Voice Builder are compatible only with Supertonic 2. Custom voice import is hidden during the v3 transition and will return when Voice Builder ships v3 support. If you previously imported custom voices, the migration removes them from Listen2; your original JSON files outside the app are unaffected.

v1.9.0 iCloud Sync Listen 2 can now keep your library in sync across all your Apple devices. Reading positions, bookmarks, voice preferences, pronunciation rules, and your documents themselves all move with you between iPhone and iPad. - Off by default. Turn it on in Settings, iCloud and Backup, iCloud Sync. - Requires iCloud Drive to be enabled in iOS Settings, with Listen 2 in the apps list. The app will warn you if iCloud Drive is off when you try to enable sync. - Sync is eventually consistent. Small changes appear in seconds, larger imports may take a few minutes. - Existing users see a one-time onboarding sheet on first launch so iCloud storage is never consumed without your say-so. - Voices downloaded on one device show a "Downloaded on another device" badge on the other, with a one-tap download button. Voice files themselves are not auto-downloaded across devices. - Full v1.9 sync and backup details in the user guide available via settings menu (link to a web based guide) Export and Import Backup A complete library snapshot you can save anywhere — Files, AirDrop, email, another cloud service. Works independently of iCloud. - Export Backup creates a .listen2backup file. Choose Full Backup (every document plus metadata) or Metadata Only (positions, bookmarks, library structure — much smaller). - Import Backup restores from a .listen2backup file. Choose Merge (keeps your existing library and adds backup contents, newer wins on duplicates) or Replace (wipes the current library and restores from backup — cannot be undone). - Designed for device transfers, Apple ID changes, and one-time safety snapshots. PDF Import PDF imports are faster and now handle scanned and image-based PDFs. - Single-pass streaming import — text PDFs land in the library noticeably quicker, especially long ones. - Automatic detection of scanned PDFs with no embedded text. When detected, Listen 2 offers to run on-device OCR (iOS 18.2 and later, using Vision). - Cancel button is available at any point during import. Redeem Code A new Redeem Code button under Restore Purchases in Settings. Opens Apple's standard offer code redemption sheet so promo codes can be applied without leaving the app. Apple now requires this surface in-app. Figure Description Prompt The "describe figures after import" prompt is now off by default for new installs. Existing users keep their current setting. Toggle it any time in Settings. Performance - Figure description on large EPUBs is dramatically faster. The image extraction step previously re-unzipped the EPUB once per image (about 5 minutes for a ~600-image book). It now opens the archive once and batch-extracts. The step is effectively instant on books that previously hung for minutes. Bug Fixes - Fixed a bug where the play button could get stuck after the app returned from the background. The audio session is now re-asserted on the foreground transition, with a self-healing fallback that catches residual cases. - Fixed a related regression where pause then resume could produce a brief blip of audio followed by silence. Audio session reactivation moved out of the resume path so it no longer wipes scheduled buffers. - Fixed a crash that could occur when resuming playback from the lock screen or Control Center. The audio player node could throw an unrecoverable error if the audio session was invalidated while the app was in the background. Resume now recovers gracefully instead of crashing. - Fixed a crash during background launches where iOS would terminate the app for holding a database file lock while suspended. The app now requests additional background execution time while loading the database, preventing the system from killing it mid-operation. - Fixed a missing guard in the audio streaming startup path that previously ignored engine initialization failures.

Reader View Layout - Some smaller devices had issues where the layout wouldn't fit on screen and thus spilled off the right side - Moved jump to current playback position and jump back buttons from top bar to the navigate button in the bottom bar to the right of the progress bar. - Moved voice picker and speed along with other audio related settings to a new audio settings button to the left of the progress bar. - The reader view quick settings sheet is now slimmed down to display related settings and it still located in the same position. Voice over - Text view now supports paragraph, line, sentence, word, and character rotors - and flicking up and down between those granularities. - Text view now supports copy paragraph, sentence or line in order to review or share text. - Now supports 2 finger z scrub to go back out of text view back to library - 2 finger swipe down for read all - Text stays on screen so that touch exploration doesn't accidentally derail text exploration too badly - Fixed issue where Voice Over labels weren't being read properly for timer preset buttons. PDF extraction improvements - Standard PDF import now supports new setting whether to filter out footnotes. Some users prefer to include the footnotes in the extracted text - Fix issue where some PDF files with large line spacing extracted each line as a separate paragraph - Fix issue where a space was inserted after commas in large numbers like 160,000 Collection Improvements - Restart Collection - already listened to a collection and want to re-listen from the beginning? Press the restart collection button and all collection document positions are reset to their beginning and you can re-listen to the entire collection start to end. - Voice Over has new actions for re-ordering collection documents in collection edit view

KOKORO VOICE ENGINE A premium neural voice engine joins Listen2. Kokoro produces more expressive, natural-sounding speech with richer intonation and smoother prosody. 53 voices across 8 languages: English, Spanish, French, Hindi, Italian, Portuguese, Vietnamese, and Chinese. All voices are included in a single engine download — tap download on any Kokoro voice and every voice becomes available. Requires devices with 6 GB RAM. Speed capped at 1.4x for real-time synthesis. LISTEN2 BUNDLED VOICE Zara, a custom-trained neural voice, now ships bundled in the app. No download needed — you are listening in seconds. Marked with a gold "Listen2" badge. Always available as a fallback if your selected voice is unavailable. TIMER Set a timer and playback pauses when it expires. Choose from 15, 30, 45, or 60 minute presets, or set any custom duration from 1 to 120 minutes. An optional audio cue plays a gentle tone when time is up. The countdown appears in the bottom bar while active. Full VoiceOver support on every control. DURATION ESTIMATION AND PROGRESS BAR Listen2 now estimates how long your document will take to read. A seekable progress bar shows your position with elapsed and remaining time. Drag the bar or enter a percentage to jump anywhere. The lock screen scrubber lets you seek without opening the app. Duration refines in real time as you listen. ENGINE BADGES Every voice in the Voice Library displays a color-coded badge — blue for Piper, purple for Kokoro, orange for Supertonic, gold for Listen2, gray for System. VoiceOver reads each badge as the engine name. VOICE LIBRARY PERFORMANCE The Voice Library loads instantly from cache, then refreshes in the background. Filter changes are debounced. Download progress updates per-voice instead of refreshing the entire list. NEW LANGUAGES (new in v1.6) Greek, Turkish, and Hindi added to the Piper engine, bringing the total to 13 languages. ACCESSIBILITY Seek-able progress bar with VoiceOver adjustable actions (5% increments). Timer controls with descriptive labels, values, and hints. Reader controls stay visible when VoiceOver is active. Engine badges announce their engine name.

- New Piper voice models and languages - Greek, Turkish and Hindi - Completely re-built highlighting engine that is leaner and more efficient and much easier on your battery life. - App bundle reduced by about 300MB

What's New in Listen2 SUPERTONIC VOICE ENGINE A second TTS engine joins Piper. Supertonic uses flow-matching neural models that we custom-converted to run natively on Apple's Neural Engine via CoreML. The result is noticeably more natural speech — smoother prosody, better rhythm, clearer articulation. Available in English, Spanish, French, and Portuguese. Download voices from the Voice Library and switch between engines at any time. Note that Supertonic voices use a fundamentally different voice architecture than Piper and can be less deterministic with their pronunciation and accuracy. If you have a custom voice from Supertonic's Voice Builder, you can import it directly from the Voice Library. Apple System Voices now available as 3rd engine option For when you want to listen but you want to preserve battery. Or when you are just used to a system voice and you prefer that one over a Piper voice. Or if you need a language that isn't supported yet with our Piper or Supertonic models. VOICE TUNING & PRESETS (Piper and Supertonic) Fine-tune any voice with quality sliders — noise scale, length scale, articulation speed, duration buffer, and more. Save your favorite settings as presets to switch between tuning profiles quickly. PRONUNCIATION RULE PREVIEW Preview how a pronunciation rule sounds before saving it. Tap the preview button to hear the replacement spoken aloud so you can dial it in without trial and error. FIGURE DESCRIPTIONS For PDF and EPUB files (more file type compatability coming), images are detected, fed through a local image AI model that generates figure descriptions. Generate brief, standard or detailed descriptions and the figure descriptions are read aloud during playback. An audio cue and haptic feedback indicates when a figure is being read. Figure descriptions automatically translate to match your document's language. When Listen2 detects a non-English document, generated figure descriptions are translated using Apple's on-device Translation framework. If the language pair is not installed, the app prompts you to download it. WEB ARTICLE COLLECTIONS Batch your loose web article imports into an auto-playing collection in two taps. Articles play back-to-back until finished. Delete the collection and its documents together when you are done. NEW LANGUAGES - Portuguese — now supported on both Piper and Supertonic engines - Custom built grammar pack for normalization across 9 categories. - Vietnamese — 3 Piper voices (one with 65 speakers) plus text normalization grammar pack SHARE EXTENSION IMPROVEMENTS Better article extraction from complex websites. Safari JS preprocessing captures authenticated page content for improved results on paywalled and Cloudflare-protected sites. Added WebView fallback for sites that resist standard extraction. PERFORMANCE Reduced memory usage ~40% during word-level highlighting. The low-level fix dropped peak memory from 1.1–1.5 GB to 600–900 MB during playback. ACCESSIBILITY - All Supertonic voice quality sliders work with VoiceOver swipe-to-adjust - Finer 0.01 increment granularity on numeric sliders - "Apply now" and "apply soon" modals are VoiceOver-visible when adjusting Piper settings - Improved VoiceOver coverage for figure description flow - Image descripton feature for PDF and EPUB files.

Web Article Import - Paste any web article URL in "Import from Link" — Listen2 extracts the article text automatically - New Safari Share Extension: tap Share → Listen2 from Safari, NYT, and other apps to save articles directly to your library Pronunciation Rules - Fix how any word is spoken — names, abbreviations, technical terms — without changing the display text - Create rules from the reader's quick settings while reading, or from main settings - Match by exact word or substring, with replace or delete output - Document-scoped rules for fixes that only apply to one document - Built-in presets for footnote markers ([1], [2], etc.) and reference symbols (†, ‡, §) - Drag to reorder, toggle on/off, full VoiceOver support Playback Fixes - Fixed AirPods Pro play/pause button not resuming playback - Fixed playback speed not saving across sessions - Fixed speed changes during playback causing audio to pause or behave unpredictably - Fixed word level highlighting falling back to uniform timing when special characters were present

Multi-Language Voice Support Download voices in Spanish, French, German, Italian, Russian, Hungarian, and Swedish — all with word-level highlighting. Our first Non-Latin script language, Russian Cyrillic, tracks correctly. Grammar packs auto-download when needed. Voice Samples & Audition Preview any voice before downloading. Over 1,100 speaker samples available across all voices. Multi-speaker models let you browse and audition individual speakers inline, with full sheets for voices with 20+ speakers. Collections Added source links to individual documents so you can let others know where the documents came from. Also added collection level description, and fixed a bug where resuming playback would get stuck in collection view. Large Document Support Completely rebuilt the document loading pipeline around SQLite-backed paragraph storage and streaming import. Large EPUBs that previously crashed now load smoothly. New Format Support — DOCX files with style resolution and heading-based table of contents — DAISY digital talking books (2.02 and ANSI/NISO Z39.86 formats) — AI Vision PDF extraction for scanned or image-based PDFs. Marked as experiemental for now. Works better for certain PDF files compared to our custom PDF extraction. Try both extractors and use keep the version that imported better. VoiceOver Improvements Line-by-line rotor navigation works now including across paragraph boundaries. Explore-to-seek: touch or rotor-navigate to any line during playback to seek there. Magic Tap reliably toggles play/pause. Highlight overlays stay visible during VoiceOver playback. Fixed an issue where scroll position didn't update when text highlighting was off. Navigation & Playback Fixes Jumping to a TOC entry, search result, or previous position now lands at the correct scroll and playback location. Playback state is preserved across navigation. Centered auto-scroll re-enabled with word-highlight gating. Seamless Paused Reading When paused, you can now scroll freely through the entire document — the text window extends seamlessly as you scroll near the edges. When playing, scrolling ahead shows a "Return to playback" floating pill to jump back Voice Tuning Adjust noise scale and noise scale W parameters per voice. Audio Quality Fixed limiter distortion with peak-aware gain and asymmetric smoothing. Prevented auto-resume after phone calls when user had paused. Settings Voice section moved near the top. Quick settings reordered: audio, reading behavior, appearance. Added contact support and rating links. Voice Library Re-organized the voice library layout for better discoverability. Users reported difficulty finding the downloadable voices so we moved those to the top and cleaned up the voice library view ensuring Voice Over compatability and speaker audition and selection functionality.

Collections let you group documents together and listen to them as a continuous playlist with smooth transitions. The reader has been rebuilt for smoother scrolling and more accurate highlighting, and now includes centered auto-scroll that keeps the spoken line pinned to the center of your screen. A new progress bar in the playback controls tracks your reading position in real time.

New Features Tap-to-Word Playback - Playback starts from double-tapped word instead of at the beginning of the paragraph. Line-Level Highlighting (Works with Word Hightlight mode only) - Current line is highlighted (configurable in settings Bookmarks System - Bookmark button in top control bar in reader view - Inline underline indicators - Long press to delete bookmark - Table of content sheet now has bookmark list. You can navigate to bookmarks or delete them from there Reader Text Display Customization - Text display customization with consistent highlight geometry - Book-style paragraph formatting Line Length Limits (for optimal landscape orientation on iPads) - CPL-based (Character Per Line) layout for improved iPad readability - Configurable in settings. Individual settings for portrait and landscape modes Sentence Highlighting - Improved sentence level highlighting Sentence Level skipping - Configure between paragraph and sentence level skipping in settings - Music-player skip behavior restored Jump to current playback position - New Crosshairs icon in reader view to jump to current playback position. Now you can scroll around and easily get back to the current line. Fixes & Improvements Reader/Playback - Multiple playback and line-level scroll bugs fixed - Scroll sync issues resolved after orientation/background changes - Scroll position syncs with playback + jump button added - Page-follow hysteresis fixed, reduced blank space - Top bar layout improved for narrow screens

Fixed playback resuming from the wrong position after leaving and returning to the app