GitHub radar
Life Recorder: Local AI Voice Diary for iPhone and Mac
Records your voice on iPhone and transcribes it locally on Mac with Whisper — no subscription, no cloud, audio never leaves your devices.
Life Recorder is a pair of apps from the browser-use team: a native iPhone app that records approximately one-minute AAC audio chunks, and a Python receiver on Mac that transcribes them locally using Whisper from whisper.cpp. The system maintains one continuous Markdown transcript file with hourly markers; audio is deleted from the phone after the Mac confirms receipt and successful transcription. No third-party transcription service or cloud backend is involved. The phone and Mac communicate over a local LAN connection secured with a self-signed TLS certificate and a 256-bit random bearer token. Building the iOS app requires Xcode and a free Apple developer account. The authors included a SKILL.md playbook so a coding agent (Codex or similar) can inspect the source, build the app, and install it on a connected iPhone with minimal manual steps.
Why a vibe-coder should care
Voice is faster than typing for capturing ideas on the go. Commercial transcription services send audio to the cloud. Life Recorder keeps everything local: no subscription, no privacy trade-off, and the Markdown transcript is immediately usable as context for any AI tool.
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Deploy life-recorder from https://github.com/browser-use/life-recorder — follow SKILL.md, build the iOS app with Xcode, and install it on my iPhone. Ask me for the Apple developer team ID and phone connection details if needed.
Requires an iPhone and a Mac with macOS. Whisper small model (~140 MB) downloaded separately. No GPU needed — transcription runs on CPU.
Open on GitHub▌ More finds