Voice Typing for Windows 11
One hotkey, every application. Outlook, Teams, Word, a browser, a PowerShell prompt, a line-of-business app from 2011 — the client types into whatever field has focus, and the same account works on your Mac and your Linux box.
Text while you speak
Streaming transcription returns words sub-second rather than after you finish, so a long paragraph is not a long silence followed by a block of text you then have to check.
Old applications included
It targets the focused field rather than integrating per app, so a Win32 tool nobody has updated since Windows 7 works the same as Outlook does.
Tray app, your own hotkey
It sits in the notification area and binds a global shortcut you choose, so it does not collide with whatever Win+ combination you already use.
Getting it running on Windows
Download and run the client
A native binary that lives in the notification area. Windows will ask for microphone access the first time it listens.
Link it to your account
The client shows a device code you confirm once in the browser. It then holds its own key — you do not sign in again on that PC.
Choose a hotkey and dictate
Pick a shortcut that does not clash with an existing Windows one, put the cursor in any text field, hold it and speak. Release, read what arrived, carry on.
You already have Win+H. When is that not enough?
Windows Voice Typing is free, built in, and genuinely usable. If it is working for you, keep it — a paid tool has to earn the difference rather than assume it.
The differences that actually decide it are narrow. Win+H is a Windows feature, so it ends at the edge of Windows: nothing about it follows you to a Mac or a Linux machine, and if you move between operating systems during a working week you are learning two or three different dictation behaviours. Behaviour also varies by application — Microsoft’s own apps are the best-supported case, and older or unusual text controls are the worst. And there is no path from it to transcribing a recording you already have; Win+H dictates, and that is all.
If all your work happens in Windows, in mainstream applications, in short bursts, Win+H may well be the right answer and this page has not made a case against it. The case is for people who cross operating systems, dictate at length, or also need batch transcription of recordings under the same subscription.
What it does not do
- It does not control Windows. No voice commands for menus, windows or the mouse — it types. Windows Speech Recognition and Talon are the tools for control.
- It needs a network connection. Recognition runs on our servers, so there is no offline mode; Win+H has an on-device path and remains the better answer on a plane.
- Some applications refuse synthetic input. Certain games, remote-desktop sessions and hardened enterprise clients block programmatic keystrokes, and no dictation tool gets text into those.
- There is no custom vocabulary. Internal product names, drug names and unusual proper nouns cannot be taught to it yet — this is the main thing Dragon still does that we do not.
Common questions
How is this different from Windows Voice Typing (Win+H)?+
Win+H is a Windows feature, so it stops at the edge of Windows and behaves best inside Microsoft’s own applications. VibeVoice is one client with one account across Windows, macOS and Linux, it streams text sub-second while you speak, and the same subscription also transcribes recordings you upload. If you work only in Windows and dictate in short bursts, Win+H is free and may be enough.
Does it work on Windows 10 as well as Windows 11?+
Yes. Nothing in the client depends on Windows 11 specifically — it binds a global hotkey and types into the focused field, which works the same on Windows 10.
Will it type into legacy or in-house applications?+
Usually, because it targets whatever text field has focus rather than integrating with each application. The exceptions are programs that deliberately reject synthetic input — some games, some remote-desktop and virtual-desktop sessions, and some hardened enterprise clients.
Does it need administrator rights to install?+
It runs as an ordinary user application in the notification area. What it does need is microphone permission, which Windows will prompt for the first time you dictate.
What does it cost?+
30 minutes a month free with no card required, then €3 a month for Pro. One account covers the Windows, macOS and Linux clients, uploaded recordings and the WhatsApp bot.
The same client runs on macOS and Linux. Coming from Dragon? That comparison is here, including what Dragon still does better.
Last reviewed . Competitor pricing and platform support change; check theirs before buying.