# Transkript — AI Transcription & Translation > AI Transcription & Translation (iPhone/iPad app by Roger Goesele.) - Source: https://appshunter.io/ios/app/transkript/id6754033691 (this page in markdown: same URL + `.md`) - Developer: [Roger Goesele](https://appshunter.io/developer/1668199673) - Category: Utilities - Price: Free - Age rating: 4+ - Requires: iOS 26.1 · 33 MB - Languages: Dutch, English, French, German - Released: 2025-11-24 - Data updated: 2026-05-23 - User reviews in markdown: https://appshunter.io/ios/app/transkript/id6754033691/reviews.md ## What is Transkript? Transform audio and video into text with Transkript – powered by Apple Intelligence for transcription, FluidAudio for speaker recognition, experimental Qwen 3 on iPad/Mac, and Gemma 4 Preview on Mac-class devices. 100% on-device privacy. SPEAKER RECOGNITION • Automatically identify who said what in your recordings • Rename speakers to real names with a simple tap • Perfect for meetings, interviews, and podcasts • Color-coded speaker labels throughout the transcript • Presets for Interview, Meeting, or Conversation scenarios APPLE INTELLIGENCE TRANSCRIPTION • Fast, accurate transcription in ~20 languages • Automatic language detection • Word-accurate timestamps for every segment • Audio enhancement to reduce background noise • 100% offline – no cloud uploads, complete privacy SESSION LIBRARY • All transcriptions organized in one place • Custom storage: iCloud, Proton Drive, or any folder • Pull-to-refresh and auto-sync • Quick access to recent sessions TRANSCRIBE ANYTHING • Audio: MP3, M4A, WAV, CAF, AIFF • Video: MP4, MOV, M4V – audio extracted automatically • Record directly in the app • Share from Voice Memos with one tap VIDEO SUBTITLES • Import videos and export SRT/VTT subtitles • Preview subtitles overlaid on your video • Auto-optimize long segments into readable captions TRANSLATE INTO 34 LANGUAGES • Translate transcripts while preserving timestamps • Export translated subtitles for international audiences AI-POWERED SUMMARIES • Generate smart summaries with Apple Intelligence, Qwen 3, or Gemma 4 Preview on supported devices • Qwen 3 is experimental on iPad/Mac; Gemma 4 Preview is Mac-class only • Visible progress when long texts are summarized in chunks • Meeting mode extracts action items and agreements • Multiple summary styles available • Export as PDF or Word (DOCX) ACCESSIBILITY • Full VoiceOver support for blind and low-vision users • Dynamic Type – text scales with your preferred size • Read Aloud: listen to summaries with text-to-speech • Subtitles for deaf and hard-of-hearing users COMPLETELY PRIVATE • All processing on-device • Optional local AI models are downloaded only when you choose to use them • No cloud uploads, no account required • One-time purchase – no subscriptions Requires iOS 26.1+ or macOS 26.1+ with Apple Intelligence enabled. ## Version history (last 5 releases) ### 4.1 — 2026-05-20 Version 4.1 – Reliability fixes FIXED • More reliable Voice Memos and share-extension handoff with unique inbox filenames • Safer summary cleanup when local models emit incomplete thinking blocks • Correct word counts after editing transcript segments • Safer WebVTT export for speaker names and special characters • More reliable cancellation and cleanup for transcription tasks • Updated FluidAudio and MLX dependencies for 4.1 • Internal release cleanup and validation improvements ### 3.9 — 2026-03-30 Version 3.9 – FluidAudio 0.13.4 & Smart Model Selection WHAT'S NEW • FluidAudio updated to 0.13.4 — dependencies reduced from 12 packages to 1 • Smart model selection: under 8 GB RAM → lightweight 110M model (~600 MB, English only); 8 GB+ → full v3 for all 25 European languages (~1.5 GB) • Language picker now limits FluidAudio to English when 110M is downloaded • One-tap switch from 110M to v3 in Settings for full multilingual support BUG FIXES • Fixed "Model assets unavailable" error on devices without Apple Intelligence (e.g. iPad 10th gen) • Deleting FluidAudio models now immediately updates the Settings screen • Added progress hint for recordings over 20 minutes with Apple Intelligence • Summarise button correctly disabled on devices without Apple Intelligence Privacy first: All processing remains 100% on-device. ### 3.8 — 2026-03-09 Version 3.8 – Stability & Accessibility BUG FIXES • FluidAudio updated to 0.12.3 with improved download progress tracking • Fixed memory leak: diarization models (~300 MB) now correctly released after transcription • Fixed enhancement UI getting stuck in loading state when enhancement fails • Cancelling a transcription no longer shows an error dialog • After app restart, model reloading now shows progress instead of a silent freeze • Fixed recording resume silently failing while UI showed it as active ACCESSIBILITY • Action bar buttons on iPhone now have proper VoiceOver labels • Recording timer, tab bar icons, and more elements now scale with Dynamic Type • Additional decorative elements correctly hidden from screen readers SECURITY • Session transcripts and audio files are now protected with iOS complete file protection, encrypted when the device is locked LOCALIZATION • Fixed status messages during model download shown in German to English, French, and Dutch users Privacy first: All processing remains 100% on-device. ### 3.7 — 2026-02-08 Version 3.7 – FluidAudio 0.12 & Liquid Glass Design FLUIDAUDIO 0.12 UPDATE • Major upgrade to FluidAudio 0.12 speaker recognition engine • Streaming audio processing for better memory efficiency • Improved speaker identification accuracy • New speaker count settings: specify 2-10 expected speakers • Orphaned word recovery: words near segment boundaries no longer lost IPHONE RELIABILITY • Fixed speaker recognition producing empty results on iPhone • Automatic fallback when word-level timing data is unavailable • Chunked audio enhancement prevents out-of-memory on long files • ML models released after use to free ~1.5 GB of memory LIQUID GLASS DESIGN • Completely redesigned with iOS 26 Liquid Glass • Beautiful glass effects with smooth morphing animations • Tab bar with fluid transitions between views • Cleaner iPad/Mac layout with streamlined toolbar KEYBOARD SHORTCUTS • ⌘, for Settings • ⌘I to Import files • ⌘R to start Recording • Perfect for iPad with keyboard and Mac ACCESSIBILITY • Improved VoiceOver support throughout • Better labels for swipe actions • Enhanced screen reader navigation Privacy first: All processing remains 100% on-device. ### 3.6 — 2025-12-14 Version 3.6 – Speaker Editing & Audio Enhancement SPEAKER NAME EDITING • Rename speakers with a simple tap – change "Speaker 1" to actual names • Names persist across the entire transcript • Makes meeting notes more readable and professional AUDIO ENHANCEMENT • New audio optimization before transcription • Reduces background noise for clearer results • Especially useful for recordings in noisy environments MEETING SUMMARIES • Enhanced meeting summary mode • Automatically extracts action items and agreements • Perfect for team meetings and project discussions IMPROVEMENTS • Faster session loading • Better error handling for translations • Improved stability throughout Privacy first: All processing remains 100% on-device. ## More apps by Roger Goesele - [Math with Tim](https://appshunter.io/ios/app/math-with-tim/id6748254953) All apps by Roger Goesele: https://appshunter.io/developer/1668199673 --- *Data collected daily from the US App Store and indexed by [AppsHunter](https://appshunter.io/). User reviews are verbatim App Store reviews. Ratings, prices and chart positions refresh continuously; this snapshot is from 2026-05-23.*