On the move — just open and speak.
Even when your hands are full — commuting, cycling, driving — with auto-start turned on (iOS 26+, supported devices), recording starts when you open it; just speak, with on-device speech recognition. Your draft is auto-saved on device, and when you send, Obsidian gets a local save — so it survives even with no signal. Opening the app and sending still take a tap; to send without touching the screen, ask Siri to “send a memo with Simple Memo” (lock settings can still require unlocking).
Holding your phone, calling, or gazing at the screen while driving or riding is illegal and dangerous in Japan (distracted cycling has been penalized since Nov 2024). Recording and sending in the app take a tap to open and a tap to send. For safety please: (1) stop in a safe place first, or (2) with a mounted phone, use only voice control such as Siri, without looking at or touching it (if it asks you to unlock or tap, stop first), or (3) have a passenger operate it. Eyes on the road first, always.
With "auto-start voice input on launch" turned on in Settings, voice input auto-starts when you open the app (iOS 26+, supported devices) — so you can start talking without finding and tapping the mic (check that it has started before you speak). What you say is auto-saved as a draft on device; when you send, it is written to Obsidian as a local file, and the email queues offline and sends automatically when you reconnect. So what you capture on the move stays on the device, even without signal.
Four things that make it work on the go
Kept on the device, even offline
In tunnels, mountains, or basements with no signal, what you are writing stays on the device as a draft, and when you send, the memo goes into the Outbox (with direct-folder Obsidian integration, it is also appended to the local note on the spot). The draft is cleared when you send. Back online, the email auto-sends and Obsidian syncs via iCloud.
Transcription stays on the device; neither audio nor text is sent anywhere for transcription. Drafts, the offline queue, and send history are AES-GCM-encrypted on device. Memo bodies are transit-only and not persistently stored on our servers.
Setup
- iPhone (iOS 26+, a supported device)
- An Obsidian vault on iCloud Drive or On My iPhone (used for local save)
- Online once (to fetch the language model)
- Auto-start voice on launch
Settings → Voice Input & Sound → "Auto-start voice input on launch." Recording then begins when you open the app. - Choose your Obsidian folder
In the Obsidian section of Settings, turn on Obsidian integration and pick your vault folder (on iCloud Drive or On My iPhone). Direct file writes mean it saves locally even offline. - Prepare the voice model (first time)
Tap the mic once while online to fetch the language model; after that, recognition works offline. - Put it on your Home Screen / widget
Place it where it's easy to reach so launching is fastest. - Open → speak → send
When it's safe, just open and talk. Sending puts the memo in the Outbox (and appends it to the local Obsidian note if integration is on); the draft is cleared.
Great for these moments
FAQ
Can I use it while driving or cycling?
Holding your phone, calling, or gazing at the screen while driving or riding is illegal and dangerous in Japan (distracted cycling penalized since Nov 2024). Recording and sending in the app take a tap to open and a tap to send. For safety: (1) stop in a safe place, or (2) with a mounted phone, use only voice control such as Siri, without looking or tapping (if it asks you to unlock or tap, stop first), or (3) have a passenger operate it. Eyes on the road first.
Does it work with no signal (tunnels, mountains)?
Yes. Offline, what you say is (1) auto-saved as a draft on device. When you send, (2) the email is queued and encrypted in the Outbox and sent automatically when you reconnect, and (3) the memo is appended to the local Obsidian note on the spot if you use folder-direct integration. Recognition also runs offline once the language model is installed.
How do I make recording start on open?
Turn on "Auto-start voice input on launch" in the Voice Input & Sound section of Settings. Recording then begins when you open the app, so you don't need to find and tap the mic (supported iOS 26+ devices; check that it has started before you speak).
How accurate is it? Any extra cost?
It uses Apple's on-device recognition (iOS 26+), and transcription stays on the device. We haven't measured accuracy; it varies with how you speak, background noise and language, so check the text before you send. Nothing is uploaded for transcription and there's no metered API, so there's no extra cost.
What if I misspeak or get interrupted?
You're covered. What you're entering is auto-saved on device (AES-GCM) about 0.4 seconds after each change, so after a close or crash you resume from the last save. Sending discards the draft.
Related Pages
References
- National Police Agency — Ban on distracted driving — Japan NPA — official info on the ban on phone use while driving
- Apple — Speech framework — Official docs for on-device speech recognition
- Simple Memo - for Obsidian — App Store — App Store download page