Free Speech to Text Online

Speech to Text: Dictate Into Your Microphone and Get Live Transcription

A speech to text tool listens to your voice through your device's microphone and turns what you say into written words on the screen as you speak, rather than requiring you to type it out yourself. Click the microphone button, start talking, and words appear and update in real time as the tool's underlying recognition engine catches up with your sentence — useful for drafting a message hands-free, capturing a spoken thought faster than you could type it, or transcribing something said out loud into text you can copy elsewhere.

This particular tool runs entirely through your browser's own built-in speech recognition feature rather than uploading audio to a transcription service you have to wait on — there's no processing delay after you finish talking, because the transcript builds up continuously while you're still speaking. It stays running through pauses, too, so a sentence broken up by thinking or a brief silence continues onto the same transcript rather than being cut off the moment you stop talking for a second.

How to Dictate and Capture a Transcript

1. Click the large circular microphone button in the center of the tool to start listening. Your browser will likely prompt you for microphone permission the first time — you'll need to allow it for the tool to hear anything.
2. Once listening starts, the button switches to a red stop icon and the status text below it changes to "Listening..." so it's obvious the tool is actively capturing audio.
3. Speak normally. As your browser's recognition engine processes what you're saying, a transcription card appears below the button and fills in with text, updating continuously as you keep talking.
4. Click the button again — now showing the stop icon — to end the listening session. The status text reverts to its resting state and the transcript stops updating.
5. Use the Copy button on the transcription card to copy the full transcript to your clipboard for pasting elsewhere, or the Clear button to erase it and start over.
6. Below the transcript itself, a running word count and character count show the size of what's been captured so far.

If your browser doesn't support speech recognition at all, the tool detects this immediately and shows a message telling you so instead of a non-functional microphone button — worth knowing before assuming something's broken if the button never appears.

What's Actually Happening: Your Browser's Speech Recognition Engine

This tool is a thin interface wrapped around the SpeechRecognition feature built into some browsers — it captures audio from your microphone and hands it to that built-in engine, which does the actual work of converting sound into text. That has real consequences for how reliably it works, depending on what browser you're in.

What this means	Why it matters
Recognition happens through the browser, not this tool	Accuracy and language support depend on the underlying engine your browser provides, not on anything this page controls
Support is inconsistent across browsers	Chromium-based browsers — Chrome, Edge, and similar — currently offer the most reliable support; others may not support it at all
Microphone permission is required	Your browser will ask to access the microphone the first time; denying that permission means the tool can't hear anything
Recognition defaults to English	The tool is set to listen for English (US) speech; other languages or strong accents may transcribe less accurately
Nothing is saved automatically	The transcript exists only in the browser tab until you copy it out — closing the tab or clicking Clear removes it with no way to recover it

If the tool shows the "not supported" message, that's a statement about your specific browser rather than a broken page — switching to a Chromium-based browser like Chrome or Edge is the most reliable fix, since browser-based speech recognition support is genuinely uneven across the market rather than universal. Firefox and Safari have historically lagged behind on this particular feature, so if dictation is something you plan to rely on regularly, it's worth confirming ahead of time that your everyday browser actually supports it rather than discovering the gap mid-task.

Situations Where Dictation Beats Typing

• Capturing a thought while your hands are busy. Cooking, walking, or doing something else with your hands makes typing impractical but leaves talking wide open as an option for getting a note or reminder down before it's forgotten.
• Drafting faster than you can type. Most people speak considerably faster than they type, so dictating a rough first draft of an email or message and cleaning it up afterward can be quicker overall than typing the whole thing from scratch.
• Reducing strain from extended typing. For anyone dealing with hand or wrist discomfort, dictating text instead of typing it removes the physical repetition that triggers or worsens that strain.
• Transcribing something said out loud nearby. A meeting note read aloud, a recipe someone's dictating from across the kitchen, or your own spoken outline for a piece of writing can all be captured directly as you or someone else speaks.
• Checking how a word or phrase actually sounds when transcribed. Speaking a name or technical term and seeing how the recognition engine interprets it is a quick, informal way to gauge how clearly you're pronouncing it.

Getting More Accurate Transcriptions

A handful of habits noticeably improve how clean the resulting text comes out. Speak at a steady, natural pace rather than rushing — recognition engines generally handle a normal conversational speed better than very fast or very slow speech, both of which can confuse where one word ends and the next begins. Reduce background noise where you can, since the microphone is picking up everything in the room, not just your voice, and competing sounds — music, other conversations, a fan or appliance running nearby — measurably increase transcription errors. Speak in complete phrases or sentences rather than single disconnected words, since the recognition engine uses surrounding context to disambiguate similarly-sounding words, and isolated single-word dictation tends to produce more mistakes than natural sentence-level speech. And review the transcript before relying on it for anything important — no speech recognition engine, browser-based or otherwise, is perfectly accurate, and homophones, unusual names, and technical terms are the most common places small errors slip in.

Reading the Transcript While It's Still Building

Because the transcript updates continuously rather than appearing all at once when you stop, it's worth understanding what you're looking at while a sentence is still being spoken. Recognition engines commonly revise a word or short phrase shortly after it first appears, as more of the sentence gives the engine additional context to reconsider an earlier guess — so text that flickers or briefly changes mid-sentence is normal behavior rather than a glitch, and it's best to treat anything still actively updating as provisional until you've finished the thought and paused. Punctuation is one of the more noticeable gaps in this kind of live dictation: unless you say words like "comma" or "period" explicitly, the recognition engine generally won't insert punctuation on its own, so a transcript of several sentences run together often needs commas and periods added back in manually after you're done. Capitalization at the start of a new sentence can be similarly inconsistent for the same reason — the engine is transcribing sound, not composing prose, so treating the raw output as a first draft to lightly edit rather than a finished piece of writing tends to produce better results than expecting publication-ready text straight out of the microphone.

What Happens to Your Voice Recording

Nothing about this tool stores or transmits an audio recording of your voice in any form you can access afterward — the microphone audio is processed by your browser's recognition engine, in real time, purely to produce the text you see. There's no audio file saved anywhere, no history of previous dictation sessions, and no account tying past transcripts to you; once you click Clear or navigate away from the page, both the audio processing and the resulting text are gone unless you copied the text out first. Whether raw audio is sent off your device at all as part of that browser-level recognition process depends on how your specific browser implements the SpeechRecognition feature — some browser recognition engines run fully on-device, while others may briefly send audio to a recognition service in order to produce a transcript, a detail controlled entirely by your browser's own implementation rather than by this tool.

Why does my browser say speech recognition isn't supported?

Browser support for the SpeechRecognition feature isn't universal. Chromium-based browsers like Chrome and Edge currently offer the most consistent support; other browsers may not implement it at all, which is what triggers this message. Switching to a supported browser is the most reliable fix.

Do I need to grant microphone permission every time I use the tool?

That depends on your browser's own permission settings rather than anything this tool controls. Most browsers remember a permission grant for a site and won't ask again on future visits, though privacy settings that reset permissions periodically can bring the prompt back.

Can this tool transcribe languages other than English?

As configured, it listens for English (US) speech. Recognition accuracy for other languages or strong regional accents depends entirely on how well your browser's underlying speech engine handles them, and results can vary noticeably from the accuracy you'd see with clear, standard English speech.

Is there an API for transcribing audio through this tool?

No. Microphone audio is captured and handed off directly to your browser's own SpeechRecognition feature — this site's servers never receive the audio or the resulting words, so there's no request-and-response cycle for an API to expose, and the feature simply isn't part of the site's metered API.

Is my microphone recording saved anywhere after I stop dictating?

No persistent recording is kept by this tool. The only output that remains is the text transcript visible on screen, and even that disappears once you clear it or leave the page unless you've copied it somewhere else first.

Why does the text sometimes change right after it first appears?

That's the recognition engine revising an earlier guess as more of your sentence gives it additional context — a normal part of how live, continuous dictation works rather than an error. Treat text as settled once you've paused and moved on to the next sentence.

Related Everyday Tools

Going the opposite direction — turning written text into spoken audio instead — the text to speech tool reads any text aloud using the same kind of built-in browser feature. Once you've got a transcript captured, drop it into the online notepad to keep editing or drafting from where the dictation left off. And for an exact word or character count on a longer transcript, the word counter gives a precise reading beyond the running estimate shown here.