Voice input · iOS 26+

Just speak. It's a memo.

On a supported iPhone running iOS 26 or later, tap the mic to turn speech into text. Transcription runs on-device without sending audio or text to an external speech-recognition server. First use needs permission and model preparation. Email delivery and Obsidian saving follow their own settings and requirements.

Download on the App Store See how to use it

By SimpleMemo Developer

In one sentence

Simple Memo voice input turns speech into text on your device. On supported iOS 26+ devices and languages, tap the mic and confirm that transcription is ready before speaking. Enabling "Auto-start voice input on launch" makes the app attempt to start voice input on ordinary launch. It does not skip first-use permissions or model preparation.

What this changes

🙌 Even hands-free
Cooking, holding a child, walking. When you can't type, capture it by voice instead.
🔒 Private, on-device
Transcription stays on the device — no audio or text sent to external speech servers. It's privacy-first by design.
💸 No metered speech API
Transcription does not use a metered external speech-recognition API. Sending limits, Premium terms, model-download data charges, and device storage are separate.

How it works

Tap the mic and recording starts; your speech becomes live text inserted at the cursor. While recording, a real waveform follows your voice level, showing "I'm listening" and "tap again to finish" without any words.

🎙️ Tap the mic → 🗣️ Speak → ✍️ Live text → 📤 One-tap send
Mix voice and typing
The keyboard stays up while recording. Dictate the gist, then fix names or numbers by hand.
Auto-stops on send
Sending while recording stops it automatically and clears the field. The "write → send → clear" rhythm is unchanged.
Tech note

Transcription uses Apple’s SpeechAnalyzer / SpeechTranscriber in iOS 26. The language model is downloaded into system storage rather than bundled with the app. It is separate from the app’s download size, but needs device storage and a connection for initial installation. If the model is not installed, wait for preparation and confirm that transcription is ready before speaking. The system may also update the model.

"Voice auto-input" — configure starting on ordinary launch

Enable "Auto-start voice input on launch" in the app’s settings to attempt to start voice input on ordinary launch. A supported device and language, microphone and speech permissions, and model preparation are required. Check the recording and transcription state before speaking. The "Voice Memo" shortcut assigned to an Action button or another trigger is a separate way to open voice capture.

Combine it

Open the app from a supported Home or Lock Screen entry, check the transcript, and send. Unlocking or authentication may be required. Time from launch to delivery or saving varies with conditions. Appending to Obsidian requires configuring the integration and vault or target on iPhone.

How to use it

  • iPhone (iOS 26 or later, on a supported device)
  • Simple Memo installed from the App Store
  • Microphone and speech-recognition permission on first use (transcription is on-device)
  1. Tap the mic (bottom-right)
    Tap the mic icon at the bottom-right of the memo screen. On first use, you'll grant mic/speech permission and prepare the language model if needed. Preparation can take time; confirm that transcription is ready before speaking.
  2. Speak
    Check that transcription is ready, then speak. Interim text appears live, and finalized words are inserted at the cursor. You can also type with the keyboard at the same time.
  3. Tap again to stop
    Tap the waveform (stop button) again to end voice input.
  4. Send
    Send with one tap as usual. If you're still recording, it stops automatically and the field clears.
To start automatically

Enable "Auto-start voice input on launch" in Settings. Confirm device and language support, permissions, and model preparation. Ordinary launch auto-start and the "Voice Memo" shortcut are separate settings.

Ways to use it

While doing chores
When your hands are busy, use your voice. Add to a shopping list or capture a to-do in a flash.
Long thoughts, fast
Speak your thinking to dump it quickly, then refine later — faster than typing it all out.
Catch ideas instantly
Idea capture is about freshness. Auto-start + voice lets you save it before it slips away.
When typing is a pain
For input-heavy languages, speaking is often faster. Dictate, then fix only what needs it.

Requirements & caveats

Requires iOS 26+ on a supported device and language

Voice input depends on Apple's new on-device speech recognition (iOS 26). Below iOS 26 this feature’s mic is hidden. Use keyboard input if your device or language is unsupported or model preparation fails.

First use may download a language model

A missing model is downloaded into system storage. Check free storage and your connection, then wait for preparation before trying transcription.

Microphone and speech permission needed

A permission dialog appears on first tap (or right at launch if auto-start is on). If you decline, you can grant it later via "Open Settings" in the alert.

FAQ

Does voice input cost extra?

Transcription does not use a metered external speech-recognition API. Memo sending limits and Premium terms are separate. Initial model installation needs a connection; data charges depend on your service.

Is my spoken audio sent anywhere?

This feature transcribes on-device without sending audio or text to an external speech-recognition server. Email sending transmits the finished memo text to the email relay and recipient. Obsidian saving and other integrations follow the selected settings and their own requirements.

What are the requirements? What if my device is not supported?

It requires a supported iOS 26+ device and language, microphone and speech permissions, and model preparation. This feature’s mic is hidden below iOS 26. Use keyboard input if your device or language is unsupported or preparation fails.

Do speech models use device storage?

Yes. Models are separate from the app’s download size and are stored in system storage. Free device storage and a connection for installing a missing model are required. The system may update models later; installation is not a promise that no further downloads occur.

What is "voice auto-input"?

Enable "Auto-start voice input on launch" in Settings to attempt to start voice input on ordinary launch. Device and language support, permissions, and model preparation are required. Check that transcription is ready before speaking. The "Voice Memo" shortcut is a separate entry.

Can I type while voice input is on?

Yes. You can combine speech and typing. Review the recognized text and correct names or misheard words before sending.

Do voice-written memos go to Obsidian too?

Voice input and Obsidian integration are separate features. Enable the integration on iPhone and configure vault or target access, then check the requirements of the chosen saving method. Email is suppressed while a supported Obsidian-only mode is active. A send action alone does not confirm delivery or saving; check the destination.

Speak, and it's a memo.
Free to download. Voice input on supported iOS 26+ devices.
Download on the App Store

Related Pages

Open Voice Memo with the Action button Obsidian Integration Apple Watch × Obsidian Obsidian Voice Input — 4 Ways Draft Auto-save Turn Your Inbox into a To-Do List Idea Capture Memo App Privacy Note to Email Voice memos with Siri & AirPods

References

Related Content