Local-firstPrivacy-firstWorks offlineBeta - Public Testing

Voice Typing

Press a hotkey, speak, and your words appear at the cursor. Local-first, privacy-first, and works offline.

AriaType dark

Type at the Speed of Thought

Voice input is 3-5x faster than typing. Speak naturally and watch your ideas flow.

Keyboard Typing

0WPM
3.75ร— faster

Voice Input

0WPM

Average typing speed vs. natural speaking speed

Voice to Text, Anywhere

Global hotkey, press to speak and release to type. Works in any macOS app, Windows in development.

Press to Speak
Release to Type
Global Hotkey
Customize trigger
Cross-App Insert
Works in any macOS app
VS Code
Slack
Notion
Chrome
Figma
Discord
Telegram
Safari
Firefox
GitHub

AI Polish, One Click

Remove fillers, formal style, concise text, AI prompts, custom rules. Turn casual speech into professional writing.

filler
Remove Fillers
1 / 5

One click to delete "um" "uh" "like", fix obvious grammar errors

Input
Um, I think, like, we can meet tomorrow?
Output
We can meet tomorrow?
formal
Formal Style
2 / 5

Turn casual speech into professional writing instantly

Input
This thing is pretty important
Output
This matter is rather significant
concise
Concise Text
3 / 5

Keep key information, remove redundancy, be concise

Input
I think we should consider this issue
Output
Should consider this issue
agent
AI Prompts
4 / 5

Generate structured prompts from voice, perfect for ChatGPT/Claude

Input
Help me write a sorting algorithm
Output
Write a sorting algorithm in Python with time complexity analysis
custom
Custom
5 / 5

Write custom instructions, customize your polish style

Input
Any input text
Output
Output by custom rules
๐Ÿ‡บ๐Ÿ‡ธ
๐Ÿ‡ฌ๐Ÿ‡ง
๐Ÿ‡จ๐Ÿ‡ณ
๐Ÿ‡ญ๐Ÿ‡ฐ
๐Ÿ‡ฏ๐Ÿ‡ต
๐Ÿ‡ฐ๐Ÿ‡ท
๐Ÿ‡ฉ๐Ÿ‡ช
๐Ÿ‡ซ๐Ÿ‡ท
๐Ÿ‡ช๐Ÿ‡ธ
๐Ÿ‡ฎ๐Ÿ‡น
๐Ÿ‡ต๐Ÿ‡น
๐Ÿ‡ท๐Ÿ‡บ
๐Ÿ‡ธ๐Ÿ‡ฆ
๐Ÿ‡ฎ๐Ÿ‡ณ
๐Ÿ‡น๐Ÿ‡ญ
๐Ÿ‡ป๐Ÿ‡ณ
๐Ÿ‡ฎ๐Ÿ‡ฉ
๐Ÿ‡ฒ๐Ÿ‡พ
๐Ÿ‡ต๐Ÿ‡ญ
๐Ÿ‡ณ๐Ÿ‡ฑ
๐Ÿ‡ต๐Ÿ‡ฑ
๐Ÿ‡น๐Ÿ‡ท
๐Ÿ‡บ๐Ÿ‡ฆ
๐Ÿ‡ฌ๐Ÿ‡ท
...

Multilingual Support

100+ languages auto-detect, Chinese optimized. Interface supports 10 languages including Chinese/English/Japanese/Korean.

100+ Languages Auto-detectSpeak any language, auto recognition
Chinese OptimizedSimplified/Traditional/Cantonese, precise dialect recognition
Multilingual Interface10 interface languages including Chinese/English/Japanese/Korean

Knows Your Field

IT/Law/Medical domain recognition enhancement. Custom vocabulary support, add specialized terms.

Add your specialized terms to make recognition understand your professional field better

IT/Tech
Security ยท Hardware ยท Software ยท Web ยท AI
Law
Civil ยท Criminal ยท Corporate ยท International
Medical
Pharmacy ยท Diagnostics ยท Cardiology ยท Neurology

100% Offline, Privacy First

100% offline processing, no network required. No account, zero data collection. Voice never leaves device.

100% Offline

Works Offline

No network required, all processing on device

No Account

Privacy Protected

Zero data collection, no registration required

Voice Privacy

Local Processing

Voice data never leaves device, absolutely secure

Free & Open Source

Completely free ยท Open Source ยท No Subscription ยท No Ads ยท No Limits

Open Source
Code is open source, contributions and audits welcome
No Subscription
Download once and use forever, no subscription fees
No Ads
Clean experience, no ads interruption
No Limits
No usage limits, no feature restrictions

How It Works

Three steps, zero friction.

1

Press Your Hotkey

Hit the global shortcut from any appโ€”browser, editor, Slack, anywhere. Recording starts instantly.

2

Speak Naturally

Talk at your normal pace. The on-device Whisper model transcribes in real time, no internet needed.

3

Text Appears at Cursor

Release the hotkey and your words are typed directly at the cursor. No copy-paste, no switching windows.

Ready to Boost Your Productivity?

Free download. No account. Works offline.

Download for macOS