Your voice. Your text.
Your machine.
Wren turns speech into text using AI that runs entirely on your Mac. No cloud. No subscription.
macOS 10.15+ · Apple Silicon & Intel · Privacy-first
Wren in action
Press a hotkey. Speak. Your words appear. That's it.
Meeting with the design team, March 15th.
We agreed to simplify the onboarding flow and reduce the number of steps from five to three. Sarah will handle the new illustrations by Friday. |
Everything you need, nothing you don't
Wren is focused on one thing: turning your voice into text, fast and privately.
100% Local
Your audio never leaves your Mac. No cloud, no servers, no data collection. Whisper runs entirely on-device.
Completely Free
No subscriptions, no tiers, no limits. Wren is free forever. Download and use it without restrictions.
Metal Accelerated
GPU-accelerated transcription on Apple Silicon. A 30-second recording transcribes in under 2 seconds.
Open Source
Every line of code is on GitHub. Inspect it, fork it, contribute to it. Full transparency.
Smart Paste
Detects if a text field is focused and pastes automatically. Otherwise, copies to your clipboard.
Personal Vocabulary
Teach Wren your names, technical terms, and corrections. It learns your words.
Choose your model
From quick notes to perfect prose. Pick the model that fits your workflow.
| Model | Size | Speed | Accuracy | Best for |
|---|---|---|---|---|
tiny.en | 78 MB | Fastest | Good | Quick notes |
base.en
Default
| 148 MB | Fast | Great | Daily dictation |
small.en | 488 MB | Moderate | Excellent | Long-form writing |
medium.en | 1.5 GB | Slower | Best | Maximum accuracy |
tiny.en base.en
Default
small.en medium.en Models download on first use and are cached locally. No internet needed after download.
Frequently asked questions
Everything you need to know about Wren.
Yes, completely. No tiers, no trials, no in-app purchases. Wren is free and open source.
Never. Whisper runs entirely on your Mac. There are zero network calls after the initial model download from Hugging Face.
macOS 10.15 (Catalina) or later. Works on both Apple Silicon and Intel Macs. Apple Silicon gets GPU acceleration via Metal.
Microphone (to hear you) and Accessibility (to paste text into your apps). Both are standard macOS permissions with clear prompts.
Yes. The default is Option+Space, but you can customize it in Settings to avoid conflicts with other apps like Alfred, Raycast, or Spotlight.
English models are included by default. Multi-language models supporting Spanish, French, German, and more are also available.
Download the .dmg file, drag Wren to your Applications folder, and launch. Grant Microphone and Accessibility permissions when prompted. That's it.
Start dictating in 30 seconds
Download Wren for free. No account needed.
macOS 10.15+ · Apple Silicon & Intel · ~148 MB for default model