AI Speech-to-Text
Purr
Purr is a free, open-source, on-device voice-to-text dictation app for macOS Apple Silicon with no cloud round-trips.
P
Purr
What is Purr?
Purr is an on-device speech-to-text dictation tool for macOS that transcribes audio locally using the Apple Neural Engine.
Purr vs Similar AI Tools
P Purr | ||||
|---|---|---|---|---|
| Pricing Model | Free | Free, Freemium | Free, Paid | Free, Freemium |
| Free Credits | ||||
| Key Features |
|
|
|
|
| Pros |
|
|
|
|
| Cons |
|
|
|
|
| Best For |
|
|
|
|
How to use Purr?
- 1Hold the hotkey (Right Option by default).
- 2Speak naturally - transcription runs entirely on-device.
- 3Release to type directly into the focused app. Enable Smart Typing for real-time word-by-word transcription.
Purr Key Features
- On-device transcription using Apple Neural Engine (no cloud round-trips)
- Zero telemetry and full privacy
- Smart Typing: words appear as you speak
- Two engines: Parakeet v2 (streaming) and WhisperKit (non-streaming)
- Custom dictionary support
- Meeting mode: captures system audio for remote participants
- MIT licensed, free and open-source
Purr Use Cases
- Hands-free dictation in any text field (Slack, Notes, Mail, Messages)
- Meeting transcription on macOS with system audio capture
- Accessibility tool for users with repetitive strain injuries
- Private, offline voice-to-text for sensitive information
Purr Pricing & Free Credits
Purr currently operates on a Free model.
This tool is completely free to use
Purr Pros & Cons
Pros
- 100% on-device, no internet needed after model download
- Free and open-source (MIT)
- Low latency with Parakeet streaming engine
- Privacy-focused: no audio leaves your Mac
Cons
- Requires macOS 14+ and Apple Silicon
- Hotkey customization limited to built-in presets
- Only English with Parakeet; other languages need WhisperKit (non-streaming)
What is Purr best for?
- macOS users seeking a private, free, and open-source dictation alternative
- Developers and privacy-conscious individuals who want on-device transcription
- Users with repetitive strain injuries needing voice input