Speech to Text
Speak instead of typing. Use live dictation to see your words appear as you talk, or record a note and have it transcribed with punctuation, then edit, copy or download the text.
- Encrypted connection
- No sign-up
- Free to use
Words appear as you speak. Your browser’s speech recognition is used: in Chrome and Edge the audio is processed by Google’s or Microsoft’s servers, in Safari by Apple.
How to use Speech to Text
- Choose the language you will speak.
- For live dictation, select Start dictation, allow the microphone and speak clearly. Select the button again to stop.
- In browsers without live dictation, choose Record, then transcribe: record up to 10 minutes and stop to get the text.
- Edit the transcript if needed, then copy or download it.
Speech to Text features
Live dictation
Words appear while you speak, with interim results shown as you go.
Thirty languages
Including English variants, Bengali, Hindi, Urdu, Arabic, Spanish, French, German, Chinese and Japanese.
Record and transcribe
A second mode that works in every modern browser and adds punctuation automatically.
Editable transcript
Correct mistakes directly in the text box; new dictation is added at the end.
Copy and download
Take the text to your document, email or notes app.
Clear about processing
The page explains which service handles your audio in each mode.
When to use Speech to Text
- Drafting emails, essays and messages faster than typing.
- Taking notes while reading or researching.
- Writing in languages whose keyboards you find slow to use.
- Reducing strain when typing is uncomfortable or impossible.
- Capturing ideas quickly on a phone.
Speech to Text FAQ
Why does live dictation not work in my browser?
Live dictation uses the browser’s own speech recognition, which Chrome, Edge and Safari provide. Firefox does not, so the tool switches to Record, then transcribe, which works everywhere.
Where is my voice processed?
In live dictation, by your browser maker: Chrome sends audio to Google, Edge to Microsoft, and Safari to Apple or processes it on the device. In record mode, the recording is sent to OpenAI through our server and deleted immediately afterwards. Nothing is stored by us.
How do I add punctuation?
Live dictation in most browsers adds little punctuation, though some recognise spoken commands like “comma” or “full stop” in certain languages. The record mode adds punctuation and capital letters automatically.
Why did dictation stop by itself?
Browsers end recognition after a pause. This tool restarts it automatically until you press stop, but a long silence or a lost connection can still end a session; just start again.
How can I improve accuracy?
Speak at a steady pace close to the microphone, in a quiet room, and choose the exact language and region you speak. A headset usually works better than a laptop microphone.
Dictation in the browser
Voice typing relies on speech recognition, which converts sound into words using acoustic and language models. Browsers expose this through the Web Speech API. A web page can ask the browser to listen, and the browser returns text as it recognises it, first as changing interim guesses and then as final results once a phrase is complete.
The recognition itself is not done by the web page. Chrome and Edge stream the audio to their makers’ cloud services, which is why live dictation needs an internet connection there, while Safari can process some languages on the device. Knowing where your audio goes matters for confidential material; for sensitive content, consider typing instead.
Recording first and transcribing afterwards uses a different approach. A complete recording gives the speech model the whole context of each sentence, so it can add punctuation, capitalisation and better choices for ambiguous words. It is slower than live dictation but usually reads better, and it works in any browser that can record audio.
Dictation is quicker than typing for most people: around 120 to 150 words a minute compared with 40 for average typists. A common workflow is to dictate a rough draft, then edit by keyboard, which combines the speed of speech with the precision of typing.