Studio Voices now download once and stay. Until now the voice model came down again after every app update — all 325 MB of it, even when nothing in it had changed. It is hosted by Apple now and kept on the iPhone from one update to the next, and the download shows its real progress as it goes.

The App Store previews now show Mi Speaks at night — the editor and the voice picker in dark mode, which is how a lot of listening actually happens. No changes to the app itself in this one.

Studio Voices now come to your Apple Watch. READ ON YOUR WATCH IN YOUR STUDIO VOICE — prepare a document on iPhone and it's sent to your Apple Watch automatically, so the watch reads it in the same Studio voice, with words highlighted as they're spoken. Documents you prepared earlier are sent too. SEE AND CONTROL THE SYNC — your library shows what's on your watch and sends in progress. Stop, retry or remove any document, or swipe it away on the watch. A LIBRARY YOU CAN MANAGE — long-press any document to see what it keeps and how much space it uses: prepared audio, its PDF, its copy on your watch. Prepare it, remove the audio, or attach or remove the PDF from the same menu. Storage now opens straight from the library.

Mi Speaks 2 is here, and it sounds completely different. STUDIO VOICES — 28 neural voices, American and British, generated right on your iPhone. No cloud, no account, works in airplane mode. You can even blend two voices into one of your own. PDF PAGE VIEW — follow along on the real page of your book, with each word highlighted exactly where it sits. Jump to any chapter. PREPARE FULL DOCUMENT — render every sentence up front so a long read plays to the end with the screen locked, in another app, or on a flight. ALWAYS KEEPS READING — with the screen off, anything not yet prepared carries on in your own system voice. Share a PDF from any app and it opens with its pages and chapters, and the lock screen's skip buttons move chapter by chapter. Reading is still free, with no length limit, and Premium now starts with a 7-day free trial.

Apple Watch speech is reliable again: sentences now play through the Watch's built-in speaker every time. Plus small refinements.

• MillerKit update: About shows the real Mi Speaks version • Support and privacy links open reliably from Help • Tappable support rows work correctly throughout the app • Behind-the-scenes reliability improvements

Now fully localized in 8 languages: Simplified Chinese, Japanese, German, French, Spanish, Korean, Brazilian Portuguese, and Italian. Speeches, playback, and every screen now read naturally in your language. New Feedback & Support: - Contact the developer directly from the app to ask a question or report an issue. - Spotted an awkward translation? Send translation feedback in one tap and help make Mi Speaks better in your language. Rating improvements: - The rate-Mi-Speaks prompt now appears at a sensible moment and no longer disables itself forever after one dismissal. Also in this release: - Reliability and performance fixes throughout.

Mi Speaks now lives on your wrist. * New Apple Watch app: listen to your speeches straight from your watch, with read-along text and word-by-word highlighting as it plays * Built-in speaker playback on Apple Watch — no headphones required * New Resume Reading widget: jump back into your last speech from the Home Screen and Smart Stack, with live word highlighting while it plays * Smoother voice selection and preview improvements * Reliability and performance fixes throughout

* Security fixes

* Improved parsing of links shared with Mi Speaks. * Improved iconography of a few buttons to be a little more intuitive.

+ Adds support for sharing content of public webpages with Mi Speaks simply by sharing the url to the content. * Fixes issues where share sheet calls would not load the shared content * Fixed issue where large voice lists prevent selection of any voice due to the list jumping back to the previously selected voice * Fixed issue where manually typing a speech was switching to the reader mode after typing a single character * Updated documentation to reflect current UI * Dropped a redundant button for switching between reader and editor modes.

+ Liquid Glass UI

* fixed issue when running the app on macOS where the current sentence in reader mode was clipped off on the left side.

* Major performance improvements made for large text being imported to the app. + Supports additional playback speeds for Premium Members ( made available through IN-App Purchases) + Added a scrub bar to quickly jump to a particular area of your document. * UI improvements to the reader mode including highlighting and easy of use improvements for reading along. + The Audio export now supports larger text and it also can cancel an in progress export. * Jumping between edit text mode and reader mode for large text is now more performant and takes you to the exact position you were at in either mode now. * The iCloud Documents folder name is now cleaner as just "Mi Speaks"

+ Speeches now sync across devices thanks to iCloud Drive. Existing speeches will automatically migrate.

* Updated app icon to support iOS 18

* Fix issue that prevented speeches shared from external apps being added to the Speeches Manager for premium subscribers.

- Fixed issue with playback progress jumping to odd percentages. - Removed the stop button since it does the same thing as pause now. - Fixed crashing issue on macOS Sonoma when running it as an iPad app on macOS.

* Fixed bugs around playback position that caused the reading to jump around the text seemingly out of order Now playback is smooth again playing from your exact cursor position and it no longer jumps to odd places.

Bug fixes and improvements

Improved backend to allow for more stable experiences when using Mi Speaks with large speeches.

Added support for using your Personal Voice for reading in app. Improved handling of large speeches with several thousand words.

Updated Privacy Policy link

Fixes a crash with speeches when speaking approaches the end of a speech.

Added a new Premium feature to save speeches along with their playback position so if you want to pause and listen to something else and pick the speech up again later your free to do so now.