Life Recorder
Life Recorder
Life Recorder quietly remembers the last hour of audio around you. A funny thing your kid said, a sudden idea, a detail from a call — tap once and it's saved, even though the moment already passed.
No planning, no pressing record at the right time. It already was.
Tap START and Life Recorder listens in the background, keeping a rolling buffer of the recent past. Old audio is continuously discarded — nothing piles up.
Something worth keeping? Save the last 5, 10 or any custom number of minutes — from the app, the notification, or the home-screen widget.
Play it with a waveform, trim to the exact seconds, share anywhere — or turn it into text with one tap.
Small app, sharp purpose — with a couple of superpowers.
The whole point: save audio AFTER it happened. Up to a full hour back, cut to the minute.
Turn any recording into text in ~100 languages. The speech model runs entirely on your phone — audio never leaves it.
See the sound, drag the handles, keep exactly the seconds that matter.
A recording schedule and skip-silence mode keep the buffer deep and the battery drain shallow.
The app is available in 30 languages, with full right-to-left support.
Recordings are never uploaded. Only what you explicitly save is kept — the rest is discarded as it ages out.
Everything stays on your device. Recordings are never uploaded, transcription runs locally, and only audio you explicitly save is kept. No account, no cloud — unless you choose to share something, it never leaves your phone.
While listening, Life Recorder keeps a rolling buffer of the last hour of audio. When you tap save, the last minutes of that buffer become a recording. Anything older than an hour is continuously and permanently discarded.
No. The buffer and your saved recordings live only on your phone. Transcription also runs entirely on-device. Audio only leaves your phone if you explicitly share it.
It's built to sip power. You can also set a recording schedule (only listen during chosen hours) and enable skip-silence so quiet stretches aren't stored at all.
About 100 spoken languages, automatically detected. The one-time speech model download (~380 MB) happens over Wi-Fi; after that transcription works offline. Automatic transcripts may contain mistakes.
Laws differ by country. In some places you must have the consent of everyone being recorded — please make sure you comply with your local laws.
Free to use. PRO unlocks unlimited listening, sharing, trimming and transcription.