Keyu: Live Lecture Captions

Keyu turns a lecture into live captions: what the teacher says appears on screen within seconds, each line followed by a translation. Recognition and translation run on the device through Apple's system frameworks, with no network round trip. Audience - Students in foreign-language lectures who need captions and translation - Attendees of international talks, seminars, and open courses - Anyone who wants bilingual notes of a class Workflow On first launch, a guided setup lets you choose languages, grant Microphone and Speech Recognition permissions, and download Apple's system recognition and translation components — the only step that requires a connection. After that, the app works offline. Tap the microphone button at the top right when class begins, and tap Stop when it ends. The session is saved automatically. Languages Recognition: English, Chinese (Mandarin and Cantonese), Japanese, Korean, Spanish, French, German, Italian, Portuguese. Translation: about 20 languages, following the device language by default. When both languages are the same, only the original captions are shown. Available languages depend on iOS. The interface is available in 14 languages. iPad Native support for full screen, Split View, Slide Over, and Stage Manager, with text size and layout adapted for the larger display. External keyboard shortcuts: ⌘R to start, ⌘P to pause or resume, ⌘. to stop. Display and after class Three text sizes; captions can show original and translation, original only, or translation only. Recording continues when you switch apps or lock the screen; the Lock Screen and Dynamic Island show elapsed time and the latest line, and you can pause and resume. Each class is saved automatically as a session you can search and review in bilingual full text. Long-press a line to copy or share it, or copy the full text. Export as TXT, SRT with timestamps, Markdown, or PDF. The audio recording stays on the device; it can be played in the app or exported as m4a. Privacy Recognition, translation, and recording all run on the device; core features rely on no server, and class content is not uploaded. The only network component is an optional, user-initiated diagnostic log upload, disclosed in the privacy policy. Caption text is included only in your encrypted iCloud system backup, if enabled. Limitations - Noise, distance, or strong accents may cause recognition errors; an orange status dot means sound is detected but temporarily not recognized - Requires iOS 26 or iPadOS 26 or later Support: [email protected]