AI Voice Recorder - YapVault

Your phone already records. It won't summarize what you said, tell you who was talking, or lock the one memo you'd rather nobody opened.

YapVault does all three on the device itself. No account, no subscription, nothing uploaded.

IT SUMMARIZES WHAT YOU SAID With Apple Intelligence, every recording gets a real name and a plain summary of what you covered — generated on the device, right after the transcript. No more scrolling a list of timestamps trying to remember which one was the contractor call. Rename anything and your name is the one that sticks.

IT KNOWS WHO WAS TALKING Turn on speaker labels and YapVault separates the voices in a conversation. Name each person once and the transcript reads like a script instead of a wall of text. An interview, a stand-up, a call with three people on speaker — you can see who said which part.

LOCK ONE MEMO, NOT THE WHOLE APP Put Face ID on a single recording. A locked memo drops out of the library and out of search until you unlock it — so the memo you don't want read is protected without making the other two hundred annoying to reach.

HALF THE SIZE An hour of audio is about 14 MB, against roughly 29 MB for the same hour at standard quality. Speech-tuned encoding, not a quality compromise — voices sound the same, the files are half as heavy. Record a whole day of meetings without watching your storage drain.

IT LEARNS THE WORDS IT GETS WRONG Long-press a misheard word to correct it. The fix joins your vocabulary and every future recording gets that name, product, or piece of jargon right — so the transcriber stops making the same mistake on the words you use most.

EVERY WORD FINDABLE Full-text search across transcripts, titles, and notes for the whole library at once. Tap a result and playback jumps straight to that moment in the audio.

TRANSCRIBED ON YOUR DEVICE Transcription starts on its own after each recording, using Apple's on-device speech model. It keeps working while the app is closed and picks up where it left off. No internet needed after the first model download.

NOTHING GOES TO A SERVER Your recordings, transcripts, summaries, and searches never leave your device. There is no account to make, no analytics, and nothing that reports back — the privacy manifest declares no collected data of any kind. The only things that ever touch the network are Apple's own: a one-time speech model download, the App Store if you leave a tip, and your iCloud backup, which is where your memos survive a lost phone.

YOURS TO TAKE WITH YOU Export every transcript as Markdown or JSON whenever you want. The audio files sit in the Files app as ordinary .m4a — no proprietary library, no lock-in.

TRIM WITHOUT LOSS Cut the dead air off either end. The trim re-encodes nothing, so the audio is bit-for-bit what you recorded, and the transcript re-times itself to match.

START IN A SECOND Record from Control Center, the Lock Screen, a home-screen long-press, or by asking Siri. A Live Activity keeps the timer in the Dynamic Island, and recording continues with the screen locked.

ALSO IN THE APP Pinned memos, written notes on any recording, list or grid layouts, playback from 0.75x to 2x, a drag-to-seek waveform, sort and filter by date, 30-day Recently Deleted with exact restore, and a full VoiceOver pass.

FREE The whole app, with no paywall, no subscription, and no ads. If it earns a place on your home screen there's a tip jar in Settings, and it unlocks nothing, because there is nothing left to unlock.

Requires iOS 26. Titles and summaries need Apple Intelligence; everything else works on any supported device.

Terms of Use (EULA): https://www.apple.com/legal/internet-services/itunes/dev/stdeula/