Your voice, typed into
Hold a key and talk: AI on your Mac turns your speech into text, right where your cursor is. Lightning fast and fully private: nothing ever leaves your Mac.
Faster than typing.
As simple as a walkie-talkie.
No windows to open, no buttons to click. Your hotkey works in every app, and the text lands right where your cursor is.
Hold your hotkey
Right ⌥ by default. Or any key or combo you like, such as ⇧F15. You'll hear a soft tick.
Speak naturally
In English, Spanish, German or any of 20+ languages. Whisper detects the language automatically.
Release. Done.
Transcribed on your Mac and pasted at your cursor. Your clipboard is left untouched.
Private by design.
Fast by default.
100% local
No cloud, no account, no API keys. Works offline, on a plane or behind a firewall.
Metal accelerated
OpenAI's Whisper model runs on your Mac's GPU, so transcription stays fast without a server.
Pastes anywhere
Slack, Mail, your editor, the terminal. If it has a cursor, it works.
Multilingual
Auto-detect, or pin one of 20+ languages from the menu bar.
Your hotkey
Any key, combo or single modifier. Set it in Settings, no extra tools needed.
Any Whisper model
From tiny to large-v3-turbo. Pick one in Settings and it downloads right there.
Sound cues
A tick when recording starts, a pop when your text lands. Always know the state.
Native & light
A tiny signed and notarized Swift app. No Electron. Updates itself from the menu bar.
Out of the way until you need it.
- Record any hotkey. Click the field, press a key or combination, and you're set.
- Switch languages in two clicks, with the active language shown in the menu bar.
- Pick your model, from tiny and fast to large and precise. Downloads happen right in Settings.
- Stay current with built-in update checks.
Free. Forever.
No catch.
Dictation apps usually come with a monthly subscription. This one runs on hardware you already own, so there's nothing to pay for.
No trial, no subscription, no account. Not even an email address.
Download for free- Unlimited dictation. No word caps, no minutes, no fair-use policy.
- Every Whisper model, from tiny to large-v3.
- 20+ languages with automatic detection.
- Nothing else to install. Everything is in the app.
- All future updates, built in.
- Fully private. No telemetry, no cloud, no tracking.
Up and running in a minute.
One download with everything inside: the app, OpenAI's Whisper speech recognition and a built-in model picker. No Homebrew, no terminal, no setup.
Install
Download for macOSDrag VoiceNoodler to Applications and open it. That's it: it lives in your menu bar from then on.
On first launch you pick a model, and macOS asks once for Microphone and Accessibility access.
Requirements
- Mac with Apple Silicon M1 or newer
- macOS 14 Sonoma or newer
Frequently asked questions
Is it really free?
Yes, free forever. No trial, no subscription, no account and no usage limits. Everything runs on hardware you already own, so there's nothing to charge for.
Does my audio ever leave my Mac?
No. Recording and transcription both run locally. The app only goes online to check for updates and to download the model you pick.
Do I need to install anything else?
No. OpenAI's Whisper speech recognition is built into the app and models download from Settings. No Homebrew, Python or anything else.
Which hotkey can I use?
Any key, a combination such as ⇧F15 or ⌃⌥Space, or a single modifier such as Right Option or Fn. Open Settings… from the menu bar, click the hotkey field and press it. Modifier hotkeys never get in the way: Right Option + E still types é.
Which languages are supported?
Whisper understands about 100 languages. The menu bar lists the 20 most common, including English, Dutch, German, French and Spanish. Leave it on Auto to detect the language automatically.
How accurate and fast is it?
That depends on the model, which you pick in Settings: smaller models are the fastest, while large-v3 gives the best accuracy.
How do I update or uninstall?
VoiceNoodler checks for updates at launch and once a day. When one is out, it asks you and installs it in one click, after verifying it comes from us. To uninstall, quit it from the menu bar and move it to the Trash. Downloaded models live in ~/.config/voicenoodler; delete that folder to remove them too.