Updated August 30, 2026 · By Sumbat.T
Talk to Text on a Computer: Every Way That Actually Works
The short answer
- Your computer already does this, for free. On Windows press Win + H. On a Mac press the microphone key, or control twice. On a Chromebook press Search + D. In every case, put your cursor in a text box first, because that is the step people miss when the shortcut appears to do nothing.
- They transcribe, and that is the whole job. You get your words with automatic punctuation. You do not get paragraphs, lists, a custom dictionary, or any way to change the shortcut on Windows.
- Browser tools like SpeechTexter and Speechnotes need no installation, which matters on a locked-down work laptop, but they type into their own editor window rather than into the app you are actually working in.
- If you dictate daily, the cleanup is the real cost. That is the gap a dedicated tool fills. BlabbyAI runs on Whisper large v3 turbo, types into any app on Windows or any field in Chrome, learns spellings you define, and reshapes what you said into an email or clean notes in the same keystroke. Free tier, no card.
Dictation that arrives finished
Talk once, and skip the cleanup
BlabbyAI types into any Windows app on Ctrl+Space and into any text field in Chrome, spells your vocabulary the way you taught it, and formats the text before it lands. 60 free credits a week, no card.
What talking to your computer actually does
Talking to text, usually called dictation, means your spoken words are converted into written text in whatever field your cursor is sitting in. You speak, the words appear, and the keyboard stays untouched. It is not the same thing as talking to an assistant, and it is not the same thing as controlling your computer by voice. Those distinctions sound pedantic until you pick the wrong tool and wonder why it will not do what you want.
Dictation types
Your speech becomes text in the field you are focused on. That is the whole job. It writes, and it does not press buttons or submit anything.
Voice control acts
Say "open Chrome" or "click Save" and the machine does it. Built for people who cannot use a keyboard or mouse, rather than for writing faster.
Windows draws this line clearly enough to be worth borrowing as a mental model. Voice typing on Win + H only dictates. Voice access on Win + Ctrl + S controls the entire system and dictates as well. Most people searching for a way to talk to their computer want the first one, get recommended the second by a forum answer, and end up confused about why it wants to number everything on their screen.
Worth knowing before you start
Almost every dictation tool, built-in or paid, needs your cursor inside a text box before it will start. It has nowhere to put the words otherwise. If a shortcut seems to be doing nothing at all, click into a text field and try it again before troubleshooting anything else.
Every method, side by side
There are three broad ways to do this: the tool built into your operating system, a website you dictate into inside a browser tab, or a dedicated application. They differ less on raw accuracy than on two things people rarely check first, namely where the text can go and whether the tool can shape it once it arrives.
| Tool | Trigger | Types into | Works offline | Formatting |
|---|---|---|---|---|
 Windows voice typing | Win + H | Any Windows text field | No, needs internet | Auto punctuation only |
 macOS Dictation | Mic key, or control twice | Any macOS text field | Yes, on-device for short passages | Auto punctuation only |
 ChromeOS Dictation | Search + D | Any ChromeOS text field | Yes, after language download | Auto punctuation only |
 SpeechTexter | Click the mic in a tab | Its own editor only | No, browser based | Spoken punctuation commands |
 Speechnotes | Click the mic in a tab | Its own editor only | No, browser based | Spoken punctuation commands |
 BlabbyAI | Ctrl + Space, rebindable | Any app on Windows, any field in Chrome | No, cloud based | Custom modes reshape the text |
The "types into" column is the one that decides most choices. A built-in tool and a desktop application will put text anywhere on the machine. A website can only ever fill its own editor, so you dictate there and then copy the result across, which is fine once and tedious forty times a day.
Windows: voice typing on Win+H
Windows 10 and Windows 11 both ship dictation, and it is the fastest route from nothing to working. Microsoft renamed it between versions, which is why people are unsure whether they have it: Windows 10 called it Dictation, Windows 11 calls the same feature voice typing. One tool, two names, same shortcut.
Click into a text box first
A document, an email body, a browser search field, anything you can type in. Windows will not open the bar without somewhere to put the text, and this is the single most common reason the shortcut looks broken.
Press Win + H
The voice typing bar appears near the bottom of the screen. On a touchscreen you can instead tap the microphone key on the on-screen touch keyboard.
Wait until it says it is listening
Speaking too early is why the first word of a sentence often goes missing, or occasionally arrives twice.
Speak at a normal pace
Do not slow down or over-enunciate. Modern recognition handles natural speech better than careful speech, and exaggerating tends to make accuracy worse rather than better.
The voice typing bar on Win+H. Free, always available, and the words arrive exactly as spoken. Source: Windows 11.
Turn on automatic punctuation while you are in there. The gear icon on the bar opens its settings, and auto punctuation adds commas and full stops from your phrasing and pauses. Without it you have to say every punctuation mark aloud, which is exhausting past a sentence or two.
The Windows limitation people hit first
Voice typing is cloud based. Microsoft documents that you need an internet connection to use it, so it stops working entirely on a plane or a dropped connection. Voice access, the other Windows speech tool, runs on your device and keeps working. If you dictate while travelling, that one line decides which tool you should use.
macOS: Dictation, and the on-device advantage
macOS has dictation built in under System Settings, then Keyboard, then Dictation. Switch it on and it gives you a shortcut, which is the microphone key on newer keyboards or pressing control twice on older machines. Put the cursor in a text field, trigger it, and speak.
The genuine advantage over Windows here is privacy and resilience. Apple silicon Macs run recognition on-device for shorter passages, so the audio does not leave the machine and dictation keeps working without a connection. If you are handling anything sensitive, that is a real difference rather than a marketing one, and it is worth knowing which side of the line your tool sits on.
A note for Mac readers on what comes next
BlabbyAI does not ship a Mac desktop application. What it does have for macOS is the Chrome extension, which runs in Chrome on macOS and covers everything you write in a browser: Gmail, Google Docs, Notion, Slack, Linear, any web form. If most of your typing happens in a browser tab, that is the whole job covered.
Chromebooks and the browser: dictation without installing anything
ChromeOS has dictation under Accessibility settings. Switch on Dictation, then press Search + D with your cursor in a text field. It downloads a language model on first use, so it works offline afterwards, which is more than Windows voice typing manages. We cover the setup and its limits in full in speech to text on a Chromebook.
This is also the category that matters if your computer is managed by an employer or a school. On a locked-down machine you often cannot install a desktop application at all, and that single constraint rules out most dictation tools before accuracy or price enters the conversation. Anything that runs inside the browser stays available: a website you dictate into, or a browser extension.
Free web tools: SpeechTexter and Speechnotes
Both are free websites you open in a tab, click a microphone, and start talking. There is nothing to install and no account needed, and both support spoken punctuation commands where you say "comma" or "new paragraph" and get the mark. SpeechTexter advertises support for over 70 languages. For a one-off block of text they do the job perfectly well.
SpeechTexter runs in a browser tab, with its spoken command list on the right. Note the advertisement rail on the left. Source: SpeechTexter.
The structural limitation is the same for both, and it is not about quality. They dictate into their own editor window and nowhere else. Every sentence you want in an email, a document, a ticket or a chat message has to be copied out and pasted across. That round trip is the entire reason people eventually stop using them: it is invisible on your first paragraph and infuriating by your fortieth.
The part nobody warns you about: the cleanup
Every tool above will get most of your words right. Speaking is roughly three times faster than typing for most people, so the raw capture is a genuine win. The reason people abandon dictation is almost never the transcription itself. It is what they have to do afterwards.
| What goes wrong | Why the built-in tools cannot fix it |
|---|---|
| Names and jargon come out wrong | General-purpose recognition has no knowledge of your vocabulary, and no built-in tool lets you teach it one. |
| The output has no shape | Auto punctuation and nothing past it. No paragraphs, no lists, no headings, no change of register. |
| The trigger is fixed | Windows will not let you move Win+H to another key. If that chord is awkward for your hands, it stays awkward. |
| Commands are a closed list | Say a phrase the vendor did not ship and it gets typed into your document as literal text instead of obeyed. |
The first row is the expensive one. A general-purpose transcriber has no idea how your colleague's surname is spelled, what your product is called, or which acronyms your industry uses. None of the built-in tools accept a custom dictionary, so a word they get wrong today they will get wrong every single day, and you will fix it by hand every single time.
What the built-in tool types
hey marcus quick update the deployment slipped to thursday because the migration needed a rollback we should still hit the sprint goal ill send the revised timeline this afternoon
What an email mode types
Hi Marcus,
Quick update: the deployment slipped to Thursday because the migration needed a rollback. We should still hit the sprint goal.
I will send the revised timeline this afternoon.
Both panels are the same sentence, spoken once. The difference is whether anything processes the transcript before it lands in the text field. That processing step is the thing built-in dictation does not have and cannot be given.
BlabbyAI: dictation that arrives already formatted
BlabbyAI is a dictation tool built around the assumption that the cleanup is the actual problem. It runs on Whisper large v3 turbo, and it comes in two forms because "computer" means different things to different people. The Windows desktop app types system-wide into any application on Ctrl + Space. The Chrome extension runs in the browser on Windows, macOS and ChromeOS alike, with nothing to install beyond the extension itself.
The Chrome extension puts a small bubble beside whatever field you are typing in, here with the language menu open. Nothing to install on the machine itself.
The piece that answers the formatting problem is custom modes. A mode is a written instruction that runs over your transcript before the text is typed. You write the instruction once in plain English, assign it a shortcut, and from then on that shortcut dictates in that shape. Blabby ships Grammar Correction, Translate to English and Email, and you can write your own.
A mode is a plain-English instruction plus a model. This one turns spoken notes into a formal email.
Because the instruction is free text rather than a fixed feature list, modes cover the spoken commands the built-in tools hard-code and restrict. Write "whenever I say new paragraph, insert a paragraph break instead of typing those words" and that is now your command. The same approach handles bullet points, deleting the last sentence, or wrapping a phrase in bold. You define the vocabulary you actually say rather than memorising someone else's list.
Custom spellings, taught once
Names, product terms and acronyms get spelled your way. Set it under Languages and it is applied on every recording afterwards, which is the gap no built-in tool fills.
A shortcut you choose
The trigger defaults to Ctrl+Space and is rebindable, including per-mode shortcuts. Windows will not let you move Win+H at all.
Shortcut settings, including a separate binding per mode. The Win+H chord in Windows cannot be reassigned at all.
On Windows there is also History, which writes every recording to your own disk the moment you stop speaking, before transcription even begins. If a transcription fails or the network drops, the audio is still there and can be re-run later, optionally through a different mode. That data stays on your machine. Blabby supports 90+ languages with auto-detect, and operates under a zero data retention policy.
What it costs
The free tier is 60 credits a week, roughly 2,000 words, with no card required. Unlimited use is $8.49 a month. See pricing for the current detail.
How to choose, honestly
Start with what you already have. If you dictate a few short messages a week, the built-in tool on your operating system is genuinely adequate and adding anything would be waste. Press Win + H, or the Mac microphone key, and get on with your day.
The picture changes when dictation becomes part of how you work rather than something you do occasionally. At an hour a day, the four problems in the cleanup table stop being annoyances and start being the bulk of the time you thought you were saving. That is the point at which a dedicated tool pays for itself, and it is worth being honest with yourself about which side of that line you are on.
Stay with the built-in tool if
You dictate occasionally, in one language, with everyday vocabulary, and you do not mind tidying the output by hand. It is free and it is already installed.
Move to a dedicated tool if
You dictate daily, your work has names and jargon that get mangled, you want the text formatted on arrival, or you need it to work in every application rather than one editor.
One more practical filter, because it overrides everything else: check what your machine will let you install. On a managed work laptop the decision is often made for you, and a browser extension is the option that survives that constraint. On your own Windows machine a desktop app reaches every application on the system, which a browser tool by definition cannot.
Add to ChromeWhen it stops working
Dictation failures cluster into a handful of causes, and they are worth checking in this order because the frequency is wildly uneven.
- Your cursor is not in a text field. By a wide margin the most common cause. The tool has nowhere to put the words, so it declines to open and looks broken.
- The connection dropped. Cloud-based dictation, including Windows voice typing, stops completely without internet. On-device tools carry on.
- The wrong microphone is selected. If the bar opens and listens but nothing appears, this is nearly always it. Check which input device is set as default, and that microphone permission is granted in privacy settings.
- You are inside a remote desktop or virtual machine. The session frequently swallows the shortcut before the host operating system ever sees it.
- The application draws its own text fields. Some older desktop software and locked-down enterprise applications do not report focus in the standard way, so dictation works everywhere else on the machine and not there.
If the problem is specific to one application, we have walked through the common ones in detail: voice typing not working, dictation in Word, and voice typing in Google Docs.
Frequently asked questions
Can you talk to your computer and it types?
Yes, and every current operating system can do it without installing anything. On Windows you press the Windows logo key and H with your cursor in a text box, and a voice typing bar appears. On a Mac you press the dictation shortcut set under Keyboard settings, which is the microphone key or control pressed twice depending on your model. On a Chromebook you enable Dictation under Accessibility and press Search plus D. All three type into whatever field your cursor is in, they are free, and they need no account. The differences that matter show up later: how accurate they are on names and jargon, whether they work without internet, and whether they can format what you said rather than transcribing it literally.
How do I talk to text on a computer for free?
Use the tool already built into your operating system, because it costs nothing and is always available. Windows 10 and 11 have voice typing on Win+H. macOS has Dictation, enabled under System Settings, Keyboard, Dictation. ChromeOS has Dictation under Accessibility, triggered with Search+D. If you would rather not use a built-in tool, browser-based options like SpeechTexter and Speechnotes run in a tab with nothing to install, though they only type into their own editor window rather than into whatever app you are using. BlabbyAI also has a free tier of 60 credits a week, roughly 2,000 words, with no card required.
What is the shortcut to talk to text on a computer?
It depends on the operating system. Windows 10 and Windows 11 use the Windows logo key plus H. macOS uses whatever you have set in System Settings under Keyboard and Dictation, which defaults to the microphone key on newer keyboards or pressing the control key twice on older ones. ChromeOS uses Search plus D once Dictation is switched on in Accessibility settings. Windows does not let you reassign Win+H to a different combination, which is a common complaint from people who find that chord awkward. Third-party dictation tools generally do let you rebind: BlabbyAI defaults to Ctrl+Space and can be changed in its shortcut settings.
Why is talk to text not working on my computer?
Work through these in order. First, make sure your cursor is actually inside a text box, because most built-in dictation tools will not open without somewhere to put the words, and this is by far the most common cause. Second, check your internet connection: Windows voice typing is cloud based and stops entirely when the connection drops. Third, confirm the right microphone is selected as the default input device and that microphone permission is granted in your privacy settings. Fourth, if you are inside a remote desktop or virtual machine session, the shortcut is often intercepted before the host operating system sees it. If the bar opens but no words appear, it is almost always microphone selection rather than the dictation feature itself.
Is talk to text on a computer accurate?
Built-in dictation is reasonably accurate on clear conversational speech in a quiet room, and noticeably weaker on proper nouns, technical vocabulary, industry jargon, and accented speech. The reason is structural rather than a bug: these are general-purpose transcribers with no knowledge of the specific names, product terms and abbreviations in your work, and no way for you to teach them. A name they get wrong today they will get wrong every day. Dedicated dictation tools compete on exactly this point. BlabbyAI runs on Whisper large v3 turbo and supports custom spellings, so you define once how a term should be written and it matches that afterwards.
Does talking to text work in any application?
The built-in tools work in most places but not everywhere. They type into whatever standard text field has focus, which covers Word, Outlook, Chrome, Slack, Notion and the large majority of desktop and web applications. Where they fail is software that draws its own text fields without reporting focus to the operating system in the expected way, which tends to mean older desktop applications, some remote desktop and virtual machine sessions, and certain locked-down enterprise software. Browser-based dictation tools have the opposite shape of limitation: they are reliable inside the browser and cannot reach anything outside it, which is why a browser extension is often the practical answer on a managed work laptop where installing software is not permitted.
What is the difference between talk to text and voice commands?
Talk to text, also called dictation, converts your speech into written words in a text field. Voice commands control the computer itself, so you can say open Chrome, click Save, or scroll down and the machine acts rather than types. Windows separates these cleanly: voice typing on Win+H only dictates, while voice access on Win+Ctrl+S controls the whole system and dictates as well. The practical difference is who each is built for. If you want to write faster, dictation is what you want. If you cannot comfortably use a keyboard or mouse at all, a voice control tool is the one designed for that. Most dictation tools, including BlabbyAI, write text and do not click buttons or submit forms.
Can talk to text add punctuation and paragraphs automatically?
Punctuation yes, structure mostly no. Windows voice typing has an automatic punctuation setting behind the gear icon on its bar, and macOS adds punctuation automatically on recent versions, both working reasonably well on straightforward sentences. What none of the built-in tools do is structural formatting. You will not get paragraph breaks, bullet lists, headings, or numbered steps, and the register of what you said is never adjusted. The output is a continuous run of punctuated sentences, which is fine for a chat message and awkward for an email or a document. Tools with an AI processing step handle this differently: a BlabbyAI mode reshapes the transcript into an email or clean bulleted notes in the same keystroke.
Is talk to text on a computer private?
It depends entirely on where the recognition happens. Cloud-based dictation sends your audio to a server, which includes Windows voice typing, and Microsoft documents that it requires an internet connection for exactly this reason. On-device recognition keeps the audio on your machine: macOS Dictation runs on-device for shorter passages on Apple silicon, and Windows voice access processes locally once its language model has downloaded. If you dictate anything confidential, that distinction is the one worth checking rather than the marketing copy. BlabbyAI is cloud based and operates under a zero data retention policy, and its History feature stores audio locally on your own machine rather than uploading it.
What is the best way to talk to text on a computer?
For occasional short messages, the built-in tool on your operating system is genuinely fine and there is no reason to add anything. People move to a dedicated tool when they hit one of four limits: accuracy on the names and terms specific to their work, no way to teach the tool a spelling, output that arrives as an unstructured wall of text needing cleanup, and no way to change the trigger shortcut. If you dictate for an hour a day, that cleanup is the real cost rather than the typing. BlabbyAI addresses all four on Windows and inside Chrome, with a free tier of 60 credits a week and unlimited use at $8.49 a month.
For more on the tools and the terminology, see our guides to dictation software and speech to text.
Try it on a real day of your own typing
60 credits every week, no card and no account setup. Use it in the apps you already work in and see how much of the cleanup disappears.