Speech to Text — Free Online Voice Dictation

Dictate with your microphone and get an editable transcript

Runs in your browser · nothing is uploaded

Your voice is handled by your browser’s speech service, not by this site

Recognition is handled by your browser maker’s service. Chrome on computers and Android sends the audio to Google’s servers, and Edge sends it to Microsoft’s online speech service. Safari, and every browser on iPhone and iPad (Chrome and Edge included), uses Apple’s speech recognition, which runs on the device or on Apple’s servers depending on the device, language and settings. Each company’s own privacy policy applies. Nothing — no audio and no text — is sent to the ToolSnack server, and the transcript stays on this page. Avoid dictating anything sensitive.

Checking browser support…

Recognition language

Transcript

Finished phrases are appended here. You can edit the text freely.

What it is

Speech to Text turns what you say into written text, right in the page. It is handy for capturing ideas during a meeting, jotting notes when your hands are busy, or talking through a first draft before you polish it at the keyboard. While you speak, the phrase that is still being worked out appears in faded italics; once the browser settles on it, the phrase is appended to the transcript box below. You can edit the transcript at any time and copy it with one click. The recognition is not performed by this site. It is done by your browser maker’s speech service: Google for Chrome on computers and Android, Microsoft for Edge, and Apple for Safari and for every browser on iPhone and iPad (Chrome and Edge there run on Apple’s engine too). Apple may process the audio on the device or on its servers, depending on the device, language and settings. Nothing is sent to ToolSnack.

How to use

Choose the recognition language. On the English page it starts as English (US).
Press Start dictation and allow microphone access when your browser asks.
Speak naturally and clearly. Unfinished phrases show up in faded text first.
Press Stop when you are done. The language is locked while dictation is running.
Correct anything in the transcript box, then press Copy.
Press Clear to empty the box and start fresh.

How it works

API: the tool uses SpeechRecognition from the Web Speech API and falls back to webkitSpeechRecognition in Chromium-based browsers. If neither exists, it says the browser is not supported.
Recognition mode: continuous and interimResults are both on. Interim results are only previewed; only final results are added to the transcript.
Joining phrases: each final phrase is trimmed and added after a single space. No space is added when the text already ends with whitespace or a line break, when the phrase starts with punctuation such as a comma or full stop, or for Japanese and Chinese, which are written without spaces. The first letter is capitalised at the very start and after a full stop, question mark or exclamation mark.
Automatic restart: if the browser ends a session while you have not pressed Stop, a new session begins. After five restarts in a row with nothing recognised, dictation stops.
Errors: blocked microphone, no speech, network failure, no microphone found and unsupported language each get their own message. “No speech” keeps dictation going; the others stop it.
Size limit: dictation stops once the transcript passes 100,000 characters. Copy it, clear the box and carry on.

Examples

Language	Final phrases, in order	Transcript
English (US)	hello there. / how are you	Hello there. How are you
English (US)	hello / , world	Hello, world
Japanese	こんにちは / 元気ですか	こんにちは元気ですか
Korean	안녕하세요 / 반갑습니다	안녕하세요 반갑습니다

Accuracy depends on your microphone, background noise, how fast you speak and the speech service behind your browser. Other Chromium browsers such as Brave or Opera often expose the feature but fail with a network error; use Chrome or Edge there. Transcripts are not saved anywhere, so copy yours before closing the page.

FAQ

Does my voice go to the ToolSnack server?

No. ToolSnack's server never receives your audio or your transcript. The recognition itself is done by your browser maker's service, though. Chrome on computers and Android sends the audio to Google's speech recognition servers and Edge sends it to Microsoft's online speech service. Safari, and every browser on iPhone and iPad including Chrome and Edge there, uses Apple's speech recognition, which runs on the device or on Apple's servers depending on the device, language and settings. Each company's own privacy policy applies, so avoid dictating passwords or other sensitive details.

Why doesn't it work in Firefox?

Dictation relies on SpeechRecognition from the Web Speech API, a draft web standard, and Firefox does not provide it in its default configuration. When the feature is missing, the start button is disabled and a message explains why. Open the page in Chrome, Edge or Safari instead.

I blocked the microphone. How do I turn it back on?

Click the site information icon at the left of the address bar, set Microphone to Allow, then press Start dictation again. While access is blocked the tool shows a message saying so. On phones the browser app itself also needs microphone permission in the system settings.

Does dictation stop when I pause?

Browsers often end a recognition session after some silence or after a while. Unless you pressed Stop, this tool starts a new session automatically. If it restarts five times in a row without recognising anything, it stops and tells you, so the microphone is not left on forever.

Does it add punctuation?

Punctuation is up to your browser's speech service. The tool joins the phrases it receives, tidies the spacing and capitalises the start of sentences. Anything missing can be fixed right in the transcript box.