Speech to text online Speak and watch the text appear.
To turn speech into text, pick your language, click the microphone and talk: the words appear as you dictate and stay on the page until you copy them or download a .txt file. The tool uses the speech recognition built into Chrome, Edge or Safari. Free, no sign-up. For a finished recording, use audio transcription instead.
Microphone off
What you say shows up here first and moves into your text when the phrase is done.
- Words
- 0
- Characters
- 0
Already have the recording? Transcribe the file with speakers and timestamps.
How to turn speech into text
- 01
Pick the language
Choose the language you're going to speak from the list. Brazilian Portuguese, English and Spanish come first, with a variant for each country. The page remembers your choice on your next visit.
- 02
Click the microphone and talk
Click start dictating and allow the microphone. Talk at a conversational pace: the words show up on the provisional text line and move into your text when the phrase is done. After a pause, listening restarts on its own until you click stop.
- 03
Fix, copy or download
The text box is editable: fix a name, add the punctuation that's missing or type in the middle of the text. Then copy it to paste anywhere, or download a .txt file. Clear deletes everything, after you confirm.
- Full guide: How to use voice typing, step by step: Windows, Mac, Word, Google Docs and your phone
- Full guide: How to transcribe audio to text, step by step: from MP3 to text with speakers and timestamps
- Full guide: How to transcribe audio in Microsoft Word: the steps, the limits and what to do when you hit them
- Full guide: How to write a podcast script: structure, template and a filled-in example
How does online speech to text work?
Speech to text, also called voice typing or dictation, means you talk into a microphone and your words become text as you speak. This page uses the speech recognition that already ships with Chrome, Edge and Safari, the feature described in the Web Speech API specification. We handle the rest: joining the phrases, adding the space between them and the capital letter after a period, a question mark or a new line, counting words and characters and keeping it all ready to copy.
While you talk, the browser returns provisional text, which can still change, and then the final text of each phrase. That's why the last few words sometimes correct themselves. The browser also stops listening after a few seconds of silence; when that happens we start it again, until you click stop. After 5 minutes without hearing anything, we stop and tell you.
- You can type and fix the text at any time, even while you're talking.
- Switch languages mid-dictation and the change applies from the next phrase.
Which browsers and languages does it work with?
It works in Chrome (on a computer and on Android), in Edge and in Safari on Mac, iPhone and iPad. In Safari, recognition runs on the same speech engine as Siri, and Siri has to be turned on in the device settings. In Firefox it depends on the version: when a browser doesn't offer speech recognition, the page says so and suggests opening it in another one.
The list has Brazilian and European Portuguese, English and Spanish with a variant for each country, and dozens of other languages, from German to Japanese. Pick the language you're going to speak: recognition doesn't detect the language on its own and can't mix two in one sentence. Which languages work varies by browser; if yours doesn't recognize the one you picked, the page tells you.
| Browser | Speech to text | Notes |
|---|---|---|
| Chrome | Yes | Computer and Android, with internet |
| Edge | Yes | Computer |
| Safari | Yes | Mac, iPhone and iPad, with Siri turned on |
| Firefox | Depends on the version | The page tells you if not |
Tips for accurate dictation
Stay close to the microphone, talk at a conversational pace and finish the sentence before you pause. Background noise, music and two people talking at once turn into wrong words. A headset with a mic usually helps more than speaking louder. To check your sound first, run the mic test.
Punctuation varies by browser and language: some add commas and periods on their own, others return unpunctuated text. When the period doesn't come, type it in the text box: after a period, a question mark or a new line, the next phrase starts with a capital letter. Names, acronyms and technical terms need a second look.
- Dictate one idea at a time and review at the end of each paragraph.
- For a new line, press Enter in the text box and keep talking.
Speech to text or transcribing a recording?
They are two different jobs. Speech to text is for writing by talking: a message, a note, a first draft, a video script. One person dictates at their own pace and fixes things right away. There's no file, no speaker separation and no timestamps.
Transcription is for what has already been recorded: an interview, a meeting, a lecture, a sermon, a podcast. For that, you upload the file. At FilmScribe, we transcribe it word by word in 90+ languages, tell the speakers apart, timestamp every line and you download TXT, PDF or SRT. Playing a recording next to the mic to dictate it usually goes badly: room noise ends up in the text and the voices blur together.
To dictate inside another program, your system has dictation too. On Windows, voice typing opens with the Windows key + H and needs an internet connection. On a Mac, Dictation is turned on in System Settings, Keyboard. This page is for when you want a clean text box, the counter and the .txt file.
| Speech to text | FilmScribe transcription | |
|---|---|---|
| Audio | From the mic, live | A file or a YouTube link |
| Speakers separated | No | Yes |
| Timestamps | No | Yes, on every line |
| Output | Text to copy or a .txt | TXT, PDF and SRT |
| Price | Free, no sign-up | 3-day trial, with a card |
More than dictation. For what's already recorded, transcription.
Frequently asked questions
How do I use speech to text on my computer?
Open this page in Chrome, Edge or Safari, pick your language, click start dictating and allow the microphone. Talk and the text appears in the box. Then copy it and paste it where you need it, or download a .txt file. There's nothing to install and no account to create.
How do I convert voice to text online for free?
To write by talking, dictate on this page: it's free with no sign-up. If the voice is already recorded in an audio or video file, use transcription instead: you upload the file and get the text with speakers and timestamps. Transcription is part of the trial, which lasts 3 days or 1 hour of media and needs a card.
Does talk to text work on a phone?
It works in Chrome on Android and in Safari on iPhone and iPad. Keep the screen on and the page open while you talk. On a phone, listening may stop at the end of each phrase; we start it again automatically until you click stop.
Why is my microphone not working or blocked?
Microphone access was denied at some point. Click the lock icon next to the address, set the microphone to allow and try again. If the browser has no speech recognition, the page says so and suggests opening it in Chrome, Edge or Safari. In Safari, also check that Siri is turned on in the device settings.
Where does the audio go when I dictate?
Your browser's speech recognition is what listens. In some browsers, like Chrome, the audio is sent to the browser's own recognition service, which is why dictation needs an internet connection. The text that comes back stays on the page until you copy or download it: close the tab first and it's gone.
How do I add punctuation with voice typing?
It varies by browser and language: some punctuate on their own, others return unpunctuated text. The safe way is to type the period, the comma or the question mark in the text box while you dictate. After a period or a new line, the next phrase starts with a capital letter.
Can I dictate a voice message or a recording to get it as text?
It's not the best route. Playing the audio next to the microphone picks up room noise and drops words. For recorded audio, send the file to transcription: we transcribe it with speakers told apart and timestamps, and you download TXT, PDF or SRT.
Sources
The official specifications this tool follows.
- 01Web Speech API: the speech recognition specification(opens in a new tab) · W3C (Web Audio Community Group)
- 02SpeechRecognition: the speech recognition interface(opens in a new tab) · MDN Web Docs
- 03New WebKit Features in Safari 14.1(opens in a new tab) · WebKit
- 04Use voice typing to talk instead of type on your PC(opens in a new tab) · Microsoft
- 05Dictate messages and documents on Mac(opens in a new tab) · Apple
Already recorded it? Transcribe the file. With speakers, timestamps and captions.
Upload the audio or video, or paste a YouTube link. We transcribe it word by word in 90+ languages, tell the speakers apart and you download TXT, PDF or SRT. Trial of 3 days or 1 hour of media.
See the result on your own recording before adding a card.