Speak freely. Transcript made on your device. Clean Markdown file lands in your git repo — branch, folder, and filename on your terms.
No server in the middle. Your notes live where your code lives.
Apple Speech runs locally. Optional cloud providers (OpenAI, Mistral) only if you bring your own key.
Each note is a Markdown commit on the branch and folder you choose. No custom sync layer.
Access token stored in iOS Keychain via expo-secure-store. Sign out wipes it.
Review transcript, tweak message, rename slug. Commit only when it reads right.
Discard after transcription or keep it. Default keeps raw audio only during the transcription step.
It's just Markdown in git. Grep it, version it, publish it — or walk away. Nothing to export.
Typography first. The transcript is the artifact.
From hitting record to a git commit, in one tap.
Tap the mic. Talk as long as you want. A live waveform follows your voice.
Audio goes through Apple Speech on-device. No audio leaves the phone unless you chose a cloud provider.
Edit the message and slug if you like, then commit. A Markdown file lands in your repo on the branch you configured.
Plain Markdown in a folder you control. Diff them, grep them, render them on your blog, or delete them. No lock-in.
# notes/coordination-challenge.md
---
date: 2026-04-20T11:32:00Z
duration: 1:21
---
Riflettevo su ogni delle all'interno
dell'autorizzazione dell'attività di
natura essendo un'attività di un...
$ git log notes/ --oneline
fb1b857 feat: coordination challenge
9aa20c1 feat: AI introduction