- Recommended podcasts: shows in your recognition language, grouped by Beginner, Intermediate and Native speed, plus quick access to shows you've played recently - Automatic recognition language: podcasts are recognized in the show's own language, so English shows no longer get furigana. You can also change the language while previewing or recognizing - If the recognition language looks wrong, the app suggests re-recognizing in the right one with a single tap - Fixed: English recognized as Japanese got furigana, and translating it into Japanese still showed English. Existing recordings are corrected automatically when opened - Fixed an issue where the subscription page sometimes showed no plans to choose

· Listen first, then transcribe: tap an episode and it starts playing right away. Tap "Recognize subtitles" when you like it — keep listening while it works, follow each step, and the player with subtitles opens where you left off. · Cloud recognition for new users: try our server's large speech model 3 times — more accurate, faster, with word-level timing. The free version plays the first 5 minutes; upgrade to Pro or Pro+ to play all of it. It doesn't use your monthly free imports, and you can turn it off in Settings. · Cloud subtitle badge: subtitles from cloud recognition are marked in the player and the library. · A tidier library and import page: "Last studied" and "Cloud" are now small icon badges, and remaining imports show as simple pips. · Podcasts open more reliably: if an episode list can't be reached, the app retries another way, and tells you clearly when a show can't be reached instead of spinning. · Other improvements and bug fixes.

New: Pro+ with cloud recognition · Your audio is recognized on our server by a large model — more accurate and faster than on-device, with word-level timing · 30 hours a month; podcast episodes are fetched by the server, so there's no upload wait · Report a wrong result in one tap; when the hours run out or you're offline, the app switches to on-device recognition · Includes everything in Pro; Pro members can upgrade right from Settings Pro members get a free monthly cloud recognition trial (turn it on from the Import tab or Settings) Subtitle editing · Word-by-word highlighting is kept when you edit a line · If a line's word highlighting is off, switch it to even highlighting Design and polish · Redesigned membership card on the Import tab; Camera Roll moved up · A clearer paywall: pick one plan, and cancel your subscription anytime · The processing screen shows only the steps that will actually run

[Free imports, now counted per import] · Free users get 3 imports per month, replacing the previous 60 minutes of recognition per month · Local files, Photos videos and Text-to-Speech are now available on the free plan · Files that need recognition can be up to 30 minutes; imports with subtitles have no length limit · ZIP import remains a Pro feature; community resources work as before [Tap a word to look it up] · Tap a word in the current sentence to see its reading, dictionary form, word class and grammar notes (e.g. ～ている, ～たら) · Open the system dictionary, or add it to Memorize with this sentence as an example · With AI set up, have AI explain what the word means in this sentence · In landscape, tap words while paused [Listening practice with Notes] · Audio-Only First Pass: listen once with only the question picture, again with subtitles, then see the explanation, in landscape playback and in exported videos · Landscape playback now plays the "Explain After Listening" text · New Explanation Index: tap an entry to jump straight to that question [Playback] · Sleep Timer: 15–90 minutes or End of Track, with a gentle fade-out · In landscape, tap the screen to play / pause (can be turned off in the "…" menu) [Memorize] · When exporting for AI completion, AI can now fill in readings for the item and its examples; readings you set yourself are never overwritten [Fixes & improvements] · Fixed a freeze and crash when opening video export options on documents with many notes · When choosing an export range, the end sentence list now starts after the selected start sentence · Community ZIP resources import in one tap; better contrast on some buttons

What's New in YomiPlay 2.2 Backup & Restore - Automatic daily incremental backups to iCloud, including subtitles, explanations, vocabulary and media - Back up manually to any folder and restore in one tap Smoother Intensive Listening - Switch between Intensive Listening and Normal Playback in one tap from the player's "…" menu - "Starred lines only" now works with loop modes - In landscape, double-tap anywhere to star the line that's playing Speaker Labels - Label who says each line (Narrator / Man / Woman, or your own names), great for listening tests and dialogues - Turn it on in Settings › Subtitle Editing Sentence Boundary Alignment - Aligns each line's start and end to the pauses in the audio, fixing clipped endings and lines that run into the next - Check a whole document or batch-process from the library; every change can be previewed, nudged or undone - Overlapping timings in newly recognized subtitles are fixed automatically Subtitle Themes - New themes: Matcha and Sakura (light), Deep Sea (dark) - Subtitle cards now match each theme; custom themes let you set the page, card and current-line colors Video Export - Exports made in light mode are light too: the preview and the exported video match what you see in the app - Optional corner badge and end card (can be turned off) - Hide the original line while its translation is read first Vocabulary - Shared vocabulary decks are now free to import (counts toward the free quota) - Quick lookup on the vocabulary home screen: add anything that's missing in one tap - AI completion can now correct readings that are easy to get wrong Also - Undo one step after deleting, splitting, merging or inserting a line - Anonymous usage statistics help us decide what to improve first. They never include your subtitles or vocabulary, and you can turn them off in Settings - Fixed speed, mode and pause settings not responding during playback - Fixed hard-to-read colors in light mode, and audio drifting ahead of the picture in Intensive Listening exports

[Notes] · Add notes to one or more subtitle lines: text, images and videos · While those lines play, the note stays in a 16:9 area, rotating through videos, images and text · In landscape, show it side by side, drag it to any corner, or fill the screen · Notes travel with .yomi / ZIP export and import, and can be filled in by AI and imported back [Better Video Export] · Burn notes into your video: question images appear in sync with the subtitles, and explanations are typed out after listening · Export just the first few minutes, a range of lines you choose, or only your starred lines [Smoother Shadowing] · Shadowing and "Speak the Answer" now use on-device system speech recognition: see your words as you speak, and it stops automatically when you finish [More Accurate Part-of-Speech] · Part-of-speech underlines are more accurate, and inflected forms like 行って can be shown as one word [My Subtitle Themes] · Save your own subtitle colors as a theme [Memorization] · Export word cards for AI to complete, then import them back · Faster and more stable with large vocabularies

【Export subtitled video】 · No more screen recording. Portrait reproduces the player page exactly — your device's screen shape, the scrolling subtitle list, word-by-word highlighting · Landscape offers one large centered line, or subtitles over the picture · Choose aspect ratio, resolution, frame rate and quality, and preview before you export · Translation speech can be baked into the video. When translations play first, the original text stays hidden until the reading ends · Long lines shrink automatically instead of running off the screen · Pro feature 【Landscape playback】 · One sentence fills the screen in landscape; with video, subtitles sit below the picture · Drag to reposition subtitles, pinch to resize 【Subtitle appearance】 · Several themes plus your own colors, set separately for portrait and landscape · Themes suited to light and dark mode · Adjust the size of furigana, romaji and translations, and underline thickness — with one tap to restore defaults 【Other improvements】 · English words, numbers and symbols can now be saved as reading preferences (2024 → にせんにじゅうよん, # → シャープ) · Share audio, video and subtitle files to YomiPlay directly from other apps · Select all podcast episodes at once when importing · Import buttons added to empty groups and the first-launch screen · Music can skip "Transcribe on split", so splitting a line no longer waits on recognition that won't succeed 【Fixes】 · Edited readings no longer revert to older content after export (affected both .yomi and ZIP) · Hand-edited readings are now used correctly for speech · Saving single-line edits is noticeably faster

Read Translations Aloud · Each line's translation is spoken after it plays, so you don't have to stop and read the subtitle · Choose whether the translation comes before or after the original, with a pause in between for thinking — a fixed length, or scaled to how long the line is · Pick the primary or secondary translation, and choose the voice Peek at a Single Translation · With translations hidden, tap the button on the current subtitle to reveal just that one line · Lines that haven't been translated yet are translated on the spot Redesigned Answer Input · The input moved under the question, showing your answer the way an answer should look · The keyboard switches to the card's answer language automatically · Long answers wrap and scroll instead of running off the edge More Flexible Review · Daily and per-deck sessions can now run new items only or reviews only, so a backlog of new items no longer blocks your reviews Also · "How do I say this?" now shows which deck it saves to before you search, and remembers your last choice · Tap the Today card on the memorization home screen to start right away · Fixed: the built-in reading dictionary and your pinned readings were ignored after manually re-splitting words

A major update to Memorize and the AI tools. MEMORIZE · New "short gloss": the one-line meaning you can read at a glance — and the one review grading checks against. System translation fills only the short gloss; AI generates the short and full definitions together. Choose which to show. · New "mark as mastered": graduate an item you already know for good, instead of waiting for the interval to grow. You can put it back into review any time. · Swipe review cards: swipe left to skip, right to go back. At the end you'll see how many you skipped, with one tap to finish them. · Examples now show furigana, and you can correct a wrong reading by hand. Separate switches for the detail page and review cards. · Turn an example into its own entry — or pull a single word out of a sentence in the reading editor. · Linked entries are now grouped by relation (Synonym: xx yy). SEARCH · Kanji and kana now match each other: searching 食べる finds an entry saved as たべる, and vice versa. Katakana and full/half-width are normalized too. · The "~" in grammar entries no longer breaks search — "~ば~ほど" finds the sentences that use it. · When adding an entry, similar existing spellings are flagged so you don't collect duplicates. AI · Rebuilt "Process whole transcript": it now only adds translations by default and no longer rewrites your hand-polished text and readings unless you ask. Every option states exactly what it overwrites. Whole-transcript translation and the background profile moved into the same panel. · Subtitle correction is now limited to actual recognition errors (homophones, wrong characters, punctuation inside a line). It no longer rewords, fixes grammar, or tidies away fillers and colloquial speech. · You can write the background profile yourself — with examples and a fill-in template. OTHER IMPROVEMENTS & FIXES · Crop the photo before text recognition for more accurate results. · Deck export/import now carries linked entries, example readings and short glosses. · Fixed duplicate star and "Done" buttons when paging between entry details. · Fixed review cards still showing the old meaning after editing it on the detail page. · Fixed the Paste menu taking several attempts to appear when long-pressing a text field.

What's New • Memorization is now free: use one deck with up to 50 items — daily learning, review, read-aloud and export all included, no subscription needed • New Notes: write up one point at a time — how a few words differ, or how a grammar pattern works. Link a note to your items and it resurfaces while you review them • Notes support AI drafting (bring your own API key in Settings), writing templates, text capture from photos, and live preview. Pasted from the web? One tap cleans up the formatting • New Cloze review: content words in sentences and passages are blanked out so you recall them from context and the translation. Each review blanks a different spot, covering the whole sentence over time • Study reminders now have separate morning and evening slots, each with its own time • Examples, explanations and notes can all be captured from photos — and one photo can add several examples at once • All 10 read-aloud voices now have names and guidance on what they suit, so you don't have to audition them one by one Stability & polish • Fixed the freeze and crash when importing large files or files stored in iCloud • Imports now check available storage first and tell you exactly how much is needed • Devices with less memory automatically use a lighter recognition model, preventing crashes during transcription • Punctuation no longer repeats itself in the romaji layer, for a cleaner three-layer view • Review cards now state whether to recall the original or the translation, and explain when you answered the wrong side instead of just showing 0% • A "+" next to "My Groups" in the library lets you create a group right there

This update fixes an important performance issue and significantly expands the Memorization experience: 【Fixes】 • Fixed the device heating up noticeably and draining battery faster during subtitle playback — the subtitle list refresh logic was rewritten to greatly cut CPU usage on the player screen • Fixed garbled titles when importing ZIPs with Japanese / Chinese filenames • Passage paragraph line breaks now display correctly 【New in Memorization】 • Set aside words you no longer want to study during review: removed from review but kept in the deck, restorable anytime • New “Type Answer” self-test: answer by typing (not just speaking) with automatic scoring • Link words and sentences as synonyms/antonyms and jump between them to study by comparison • Photo capture now supports checking multiple lines, merging into a passage, and editing each entry • Mark items as starred and filter by them; AI-translate into all target languages in one tap from the detail page Thank you for using YomiPlay!

This update brings a major upgrade to the Memorization module, plus playback and speech improvements: 【Decks】 • Deck Sprint: start studying any deck directly, free from the daily plan limit • AI Batch Generate: create a batch of words/sentences by theme (with readings & meanings) — the deck name is the default theme • Reorder decks by drag & drop; customize monogram text and color • Speak-to-Answer: answer aloud during review — scored by pronunciation, and auto-marked as “Good” when you pass • Secondary translation: store translations in two languages and use the second language as your recall prompt • More powerful item details: fix item types, switch translation languages, remember or pin preferred readings 【Playback & Speech】 • Key-line marking: swipe right to star lines, then “Play Key Lines Only” to skip the rest and save listening time • Re-recognize subtitles for any audio/video; new “AI Translate All” fills in translations in one tap • More natural TTS: sentences are read from the original text, with a selectable voice remembered globally Plus many refinements and performance improvements. Thank you for using YomiPlay!

Added Memory Books. Save words, phrases, and sentences from subtitles, generate audio automatically, and create daily study and review plans. Memory Books can be exported, shared, and imported with one tap. You can also configure your own AI API key to refine subtitles, add readings and translations, and generate vocabulary, definitions, examples, and themed study content.

Fixed an issue where speech recognition failed without a network connection. Imported local audio can now be recognized fully offline.

This update introduces a brand-new Text-to-Speech feature along with multiple experience improvements. New: Text-to-Speech • Paste any text to generate narrated audio with synchronized subtitles • Supports Japanese / English / Korean / Vietnamese (on-device AI) and Chinese (iOS system voice) • Preview 10 built-in voices to pick your favorite • Automatic sentence splitting + furigana generation included More app languages • 7 new interface languages Easier to start • Built-in sample "YomiPlayへようこそ" — try the player right after first launch • New folder long-press hint to discover rename / delete / export as ZIP Other • Added "My Furigana Dictionary" — override default readings for kanji you care about • Fixed SRT subtitle import compatibility • Polished various UI details

Edit and refine Japanese furigana (reading hints) on subtitles to match your study preferences. Fixes and improvements for video import and export reliability. Minor polish and stability improvements.

What’s New in 1.1.1 • Easier to discover subtitle editing—long-press a line in the player (with a one-time tip). • Visit the YomiPlay Community from Help to share and download .yomi subtitles. • Help and other in-app text improvements. • Bug fixes and small UI polish.

Smarter translation: Batch translate subtitles without confusing double alerts. Language pack hints appear mainly when you change the translation target—which matches how the system language already works by default. Podcast & URL import: Search history, clearer hints, bottom-right URL drawer, and all episodes stay visible—over-free-limit episodes prompt Pro when you try to import. Recognition stability: Defaults favor the Small model for faster, more reliable first runs; optional heavier models on capable devices, with automatic fallback on some failures. Playback & polish: Finer speeds (0.8× / 0.9×), updated help/FAQ, and clearer notes for translation OS requirements and link import limits.