• A model download paused for a long time now starts again instead of failing, and cancelling one frees its space at once • Audio files whose name doesn’t match their format (an “.mp3” that is really AAC) now import • A recording started right after you stop or discard another no longer comes out empty • Clear all data now also erases your custom vocabulary, your usage totals and the copies kept of imported files • Fixes for rare cases where an edit, an imported backup or audio not yet transcribed could be lost (a full phone, data that couldn’t be read) • No ad or rating prompt over an open transcript or sheet, or as a recording starts • Subtitles no longer overlap, the time left stays accurate when you switch apps, and the Models tab tells every model apart • Dates stay right after you change time zone, search is quicker, and the app frees its memory when you switch away • As always: every word is transcribed on your device, never uploaded

• Private Transcribe now speaks Spanish, Portuguese, German, French, Japanese, Korean and Chinese: the whole app follows your device’s language • Watch the words appear while your recording is transcribed • Find in a transcript: every match is highlighted, and the arrows step through them • Skip back or ahead 15 seconds, or tap the progress bar to jump; the line being played stays in view • Hold a timestamped line to copy just that line • Nothing lost by accident: discarding a recording, leaving the Record tab mid-recording and closing unsaved edits now ask first, and a phone running out of space stops the recording and keeps it • Cleaner transcripts: notes like “[Music]” are left out, and a line repeated over and over appears once • On a device set to a language other than English, the first download is now the multilingual model, so your own language transcribes right away • Model downloads say when they’re waiting for a connection, and offer to switch to the new model when done • Clearer messages when a purchase waits for approval, was made before, or can’t reach the store • Search ignores accents, Japanese and Chinese transcripts show a real word count, and text is easier to read • As always: every word is transcribed on your device, never uploaded

• Open files from other apps: send an audio or video file to Private Transcribe from the share sheet or the Files app • Titles: an imported file keeps its name as the title, and you can rename any transcription; search finds titles too • Play along with a transcript: tap any line to jump there, follow the highlighted line as it plays, and change speed from 0.75× to 2× • Your edits now carry into timestamps and subtitles, and Pro adds WebVTT (.vtt) export • Nothing gets lost: recordings you haven't transcribed wait in a list, every unfinished recording is recovered if the app closes, and long histories are stored safely • Fixed: videos from the iPhone camera (.MOV) now import • Search shows where your words appear, long transcriptions show the time left, and downloaded models no longer fill your iCloud backup • As always: every word is transcribed on your device, never uploaded

• Import audio and video: transcribe voice memos, meeting recordings, lecture videos and podcasts (m4a, mp3, wav, aac, mp4, mov) — decoded and transcribed on your phone • Smaller downloads: Accurate is now 190 MB and Professional 539 MB with the same accuracy, and a new Best tier (Whisper large-v3-turbo, 574 MB) for top accuracy • Long recordings: pause and resume, keep recording when the screen locks or you switch apps, and recover an unfinished recording if the app is closed • Custom vocabulary: names and jargon come out spelled your way • Timestamps for every line, a spoken-language picker, cancel for long transcriptions, backup and restore of all transcripts • Fresh design with a new icon, models grouped into clear tiers, and downloads you can pause or cancel • As always: every word is transcribed on your device, never uploaded

• Pro (one-time purchase): remove all ads, translate 99+ languages into English text, export .txt and .srt subtitle files • Word counts on every transcript and optional auto-copy to clipboard • Storage check before large model downloads • Screen stays awake through long transcriptions • Updated Whisper engine and latest app runtime • As always: every word is processed on your device, never uploaded

bug fixes

fix audio input