Push-to-talk hotkey
Ctrl+Win works globally, so dictation starts where you already work.
Windows push-to-talk dictation
Voice Flow turns speech into text in any Windows app.
Hold Ctrl+Win, say a phrase, release the keys, and the recognized text is pasted into Notepad, your browser, chat, IDE, or any focused input.
- Free · No ads
- Windows 10 / 11, 64-bit
- Local faster-whisper first
- 90+ recognition languages
- MIT licensed
Built for fast dictation without switching windows
Voice Flow runs from the system tray and keeps the workflow simple: one global hotkey, microphone capture, speech recognition, and automatic paste into the active window.
Voice Flow 1.2.9
Website languages: English, Spanish, French, German, Portuguese, Chinese, Arabic. The app can recognize 90+ Whisper languages.
Key features
Active-window paste
No separate transcription box and no manual copy-paste.
Offline-first recognition
faster-whisper runs locally by default; Google is only a fallback.
Auto or fixed language
Use automatic speech language detection or lock recognition to one language.
Spoken punctuation
Say period, comma, question mark, and other punctuation words.
Tray controls and logs
Settings and logs stay in %USERPROFILE%\.voice-flow\.
How to use Voice Flow
- Download and run VoiceFlow-1.2.9.exe.
- Allow microphone access when Windows asks.
- Focus the window where you want the text.
- Hold Ctrl+Win, speak, then release the keys.
- Choose Automatic or a fixed language from the tray menu.
Windows SmartScreen notice
Windows may show "Windows protected your PC" because the exe is unsigned. Choose More info -> Run anyway, and verify the SHA-256 if needed.
System requirements
- OS
- Windows 10 or 11, 64-bit
- Microphone
- Required
- RAM
- 4 GB minimum, 8+ GB recommended
- Disk
- ~500 MB for the Whisper model
- Internet
- Optional after the model is downloaded
- Permissions
- Regular user account
Privacy
- Primary recognition runs on your computer.
- Voice data is not sent to dev4rweb servers.
- Settings and logs stay in %USERPROFILE%\.voice-flow\.
- If Google fallback is used, Google policies apply.
Recognition languages
Voice Flow uses Whisper for speech-to-text. You can lock one language in the tray menu or choose Automatic to let the app detect the language.
Show all recognition languages (99)
Language names are shown in their native form, as in the app tray menu.
- Afrikaans
- Azərbaycan
- Bahasa Melayu
- Basa Jawa
- Basa Sunda
- Bosanski
- Brezhoneg
- Català
- ChiShona
- Cymraeg
- Dansk
- Deutsch
- Eesti
- English
- Español
- Euskara
- Français
- Føroyskt
- Galego
- Hausa
- Hrvatski
- Indonesia
- Italiano
- Kiswahili
- Kreyòl ayisyen
- Latina
- Latviešu
- Lietuvių
- Lingála
- Lëtzebuergesch
- Magyar
- Malagasy
- Malti
- Nederlands
- Norsk
- Norsk nynorsk
- Occitan
- Oʻzbek
- Polski
- Português
- Română
- Shqip
- Slovenčina
- Slovenščina
- Soomaali
- Suomi
- Svenska
- Tagalog
- Te Reo Māori
- Tiếng Việt
- Türkmen
- Türkçe
- Yorùbá
- Íslenska
- Čeština
- ʻŌlelo Hawaiʻi
- Ελληνικά
- Башҡорт
- Беларуская
- Български
- Македонски
- Монгол
- Русский
- Српски
- Татар
- Тоҷикӣ
- Українська
- Қазақ
- Հայերեն
- ייִדיש
- עברית
- اردو
- العربية
- سنڌي
- فارسی
- پښتو
- नेपाली
- मराठी
- संस्कृतम्
- हिन्दी
- অসমীয়া
- বাংলা
- ਪੰਜਾਬੀ
- ગુજરાતી
- தமிழ்
- తెలుగు
- ಕನ್ನಡ
- മലയാളം
- සිංහල
- ไทย
- ລາວ
- བོད་ཡིག
- မြန်မာ
- ქართული
- አማርኛ
- ខ្មែរ
- 中文
- 日本語
- 한국어
FAQ
Is Voice Flow free?
Yes. The download is free, the app has no ads, and the license is MIT.
Does Voice Flow work offline?
Yes, after the local faster-whisper model is available.
Why does SmartScreen warn me?
The exe is not code-signed yet; the source code and SHA-256 are public.
What languages can it recognize?
Whisper supports 99 speech languages, plus automatic detection. See the full list in the Recognition languages section above.