To use speech to text on Windows 11, click in any text box and press Win+H. Voice typing opens, listens through your microphone and types what you say. Say "period" or "new line" for punctuation and "stop listening" to finish. It needs an internet connection because it uses Microsoft's online speech recognition.
How do you turn on voice typing in Windows 11?
- Put the cursor where text should go. Any text field works: Word, Notepad, a browser form, a chat box.
- Press Win + H. A small voice typing bar appears with a microphone button. On a touch screen, tap the microphone on the touch keyboard instead.
- Allow microphone access the first time, if Windows asks.
- Speak normally. Text appears as you talk. Pausing for a second is fine.
- Stop. Say "stop listening", click the microphone button, or press Win + H again.
To move around the voice typing bar with the keyboard, press Win + Alt + H. On Windows 10 the same Win + H shortcut opens the older dictation toolbar, which supports fewer languages and only takes commands in US English.
Settings worth turning on
Click the gear icon on the voice typing bar. Microsoft lists these options (Microsoft Support: voice typing):
| Setting | What it does | Turn it on? |
|---|---|---|
| Automatic punctuation | Inserts commas and periods based on how you speak | Yes, for prose. Saves saying "comma" fifty times |
| Voice typing launcher | Shows a mic button when you click into a text box | If you use it daily |
| Default microphone | Picks which mic voice typing listens to | Set it if you have a headset and a webcam mic |
| Profanity filter | Masks swear words | Your call |
| Wait time before a command runs | How long it pauses before acting on a spoken command | Raise it if commands fire by mistake |
What can you say to add punctuation and edit text?
Voice typing understands punctuation names and a short list of editing commands. The exact list depends on your language. These work in English:
| Say | Result |
|---|---|
| "period", "comma", "question mark", "exclamation mark" | The punctuation mark |
| "new line" | Line break |
| "new paragraph" | Paragraph break |
| "delete that" | Removes the last thing you said |
| "select that" | Selects the last thing you said |
| "undo that" | Undoes the last change |
| "stop listening" | Turns voice typing off |
The habit that makes dictation faster: talk in full sentences, then fix mistakes with the keyboard afterwards. Correcting each word by voice as you go is where people give up.
How do you change the voice typing language?
Voice typing listens in your current input language, the one shown at the right of the taskbar. To add another:
- Open Settings > Time & language > Language & region.
- Click Add a language, pick it, and make sure speech recognition is included in the optional features.
- Switch input languages with Win + Space, then press Win + H.
Windows 11 voice typing covers far more languages than Windows 10 dictation did, including several English variants (US, UK, India, Canada, Australia, New Zealand), Hindi and Japanese. The spoken command list varies by language, so "delete that" may not work outside English.
Does Windows speech to text work offline?
Voice typing doesn't. Microsoft says it uses online speech recognition powered by Azure Speech services, so your audio goes to Microsoft and you need an internet connection. There are three ways around that:
- Fluid dictation on Copilot+ PCs. On those laptops, an on-device small language model cleans up grammar, punctuation and filler words. It's on by default, English-only, and switches off in password fields.
- Voice access. Windows 11's full voice control feature (Settings > Accessibility > Speech) runs speech recognition on the device after you download its language files. It's built for controlling the whole PC by voice, so it's heavier than voice typing for plain dictation.
- Whisper running locally. OpenAI's open-source speech model runs entirely on your PC. Accuracy is very good, but it transcribes after you finish speaking rather than live. Our guide to running Whisper locally on Windows covers setup.
The old Windows Speech Recognition app, which dates back to Windows Vista, is deprecated, and Microsoft points users to voice access instead.
What are the other speech to text options on Windows?
| Tool | Cost | Works offline | Where it types | Good for |
|---|---|---|---|---|
| Windows voice typing (Win+H) | Free | No | Any text box | Everyday dictation |
| Voice access | Free | Yes | Any app, plus PC control | Hands-free use, accessibility |
| Microsoft Word Dictate | Microsoft 365 plan | No | Word documents | Long documents |
| Google Docs voice typing | Free | No | Google Docs in Chrome | Docs users |
| Whisper (local) | Free, open source | Yes | Files or apps that embed it | Recordings, privacy |
| AuraCloak voice input | Included with AuraCloak | With Local Whisper | AI chat prompts | Talking to ChatGPT, Claude or Gemini |
Why is voice typing not working?
Most failures fall into a handful of causes. Check them in this order:
- No text box is focused. Win+H does nothing on the desktop or in a window with no editable field. Click into one first.
- Online speech recognition is off. Go to Settings > Privacy & security > Speech and turn on Online speech recognition. Voice typing depends on it.
- The app can't use the mic. Settings > Privacy & security > Microphone. Make sure microphone access and "Let desktop apps access your microphone" are on.
- Wrong microphone. Pick the right one in the voice typing gear menu. Webcam mics on the far side of a desk produce a lot of misheard words.
- Language mismatch. Voice typing follows your input language. If you switched keyboards to another language, it listens for that one.
- No internet. Voice typing shows an error when it can't reach Microsoft. Use voice access or local Whisper offline.
Talk to ChatGPT, Claude or Gemini with Alt+V, including an offline Whisper engine. AuraCloak adds voice to an always-on-top AI overlay.
Get AuraCloakCan you use speech to text for AI prompts?
Yes, and it's one of the better uses for it. Spoken prompts tend to be longer and more specific than typed ones, because talking is cheaper than typing. Win+H works in the ChatGPT and Claude apps and in their web pages. The friction is that you dictate, check the text, then press Enter, every time.
AuraCloak's voice input skips those steps. Press Alt + V, speak, and it transcribes and sends the prompt when you stop. Alt + Shift + V cancels. It's push-to-talk from your own microphone only; it doesn't listen to call audio. You pick the engine in Settings > Voice:
| Engine | Cost | Audio leaves your PC? |
|---|---|---|
| Windows built-in (default) | Free | No |
| Local Whisper (base 141 MB or small 465 MB model) | Free | No |
| OpenAI Whisper API | About $0.006 per minute | Yes, to OpenAI |
| Google Speech-to-Text | About $0.009 per minute | Yes, to Google |
| Azure Speech | Billed per audio minute | Yes, to Azure |
The answers land in an overlay that stays on top and is left out of screen shares. Load ChatGPT or Claude.ai with your own login in browser mode, or use your API keys with BYOK. $9 a month or $49 once on the pricing page.
FAQ
What is the shortcut for speech to text in Windows 11?
Win+H. Click into any text box first, press Win+H, and voice typing starts listening. Press it again or say "stop listening" to finish. Win+Alt+H lets you move around the voice typing menu with the keyboard.
Is Windows voice typing free?
Yes. Voice typing is built into Windows 11 and Windows 10 at no cost. It needs a microphone and an internet connection, because it uses Microsoft's online speech recognition.
Can Windows 11 do speech to text offline?
Not with voice typing. Voice access, Windows 11's voice control feature, recognizes speech on the device after downloading language files. Open-source Whisper also runs fully offline on your PC.
How do I add punctuation with voice typing?
Say the punctuation name, such as "comma", "period" or "question mark", or say "new line" and "new paragraph". You can also turn on automatic punctuation from the voice typing settings gear.
Does voice typing work in every app?
It works in almost any app with a text field, including browsers, Office, Notepad and chat apps. It turns off in password fields. Some games and remote desktop windows don't accept it.
Why does Win+H do nothing?
Usually no text box is focused, or online speech recognition is turned off. Click into a text field, then check Settings, Privacy and security, Speech, and make sure online speech recognition and microphone access are on.