How this speech to text tool works
The tool uses speech recognition built into your browser, part of the Web Speech API. It is described in the Web Speech API specification, a draft report of a W3C Community Group (this page follows the edition of 18 September 2026). The page asks the browser to listen and transcribe what you say, and writes the text into the box. Freebiefy runs no speech server: there is no account, no upload to us and no limit on how long you dictate.
While you talk, the browser sends first guesses that can still change. They appear in gray under the text. As soon as the browser settles on a phrase, it moves into the box as normal text, with a capital letter at the start of each sentence. The box stays editable the whole time, so you can fix a word, add a heading or paste something in while the dictation goes on. New words are always added at the end.
Browsers end a recognition session by themselves, for example when you stay quiet for a while. With Keep listening switched on, the tool starts a new session as soon as one ends, so long dictation and voice typing keep working. On Android, MDN's compatibility data notes that Chrome ignores the setting that asks for continuous recognition, so on a phone the restart is what keeps the microphone open.
Where your voice goes
Speech recognition is done by your browser, and the specification allows both server-based and on-device recognition. In practice that means:
- Chrome sends the audio to Google. MDN notes that in Chrome, recognition on a web page uses a server-based engine, so your audio is sent to a web service and dictation does not work offline. Chrome's own source code describes the request: audio is sent to Google's servers and text comes back.
- On-device recognition. The specification adds a processLocally setting that requires recognition on your device. MDN's compatibility data lists it in desktop Chrome from version 139, but not in Chrome for Android or Safari. When your browser offers it, this page shows Only on this device. Tick it and the tool checks whether the language is ready; if your browser needs a language pack first, it downloads it when you press Start. If the browser cannot confirm on-device recognition, the tool does not start listening and tells you why. Microsoft's Edge documentation says the speech input of its on-device model never leaves the device.
- Safari does not document where it processes speech for web pages. Apple's Siri, Dictation and Privacy page says that dictation is sent to and processed on Apple's servers unless your device shows on-device processing in its keyboard settings. Assume the audio may leave your device.
Freebiefy receives none of it. The text stays in this browser tab and is not saved unless you tick Keep my text on this device, which stores it only in your own browser; untick the box to delete it. The specification also requires browsers to ask for your consent before listening and to show clearly when audio is being recorded, which is why your browser asks before the first dictation and shows while it is listening.
Voice commands for punctuation
When the language is English, you can speak punctuation. The tool replaces these words with the mark, removes the space in front of it and starts the next sentence with a capital letter:
| Say | You get |
|---|---|
| period or full stop | . |
| comma | , |
| question mark | ? |
| exclamation mark or exclamation point | ! |
| new line | starts a new line |
| new paragraph | leaves an empty line |
The commands are whole words, so "periodic" or "commander" stay as they are. If you need the word itself, untick Voice punctuation for a moment. For other languages the commands are switched off, because they are English words, and the text arrives exactly as the browser recognizes it. Any punctuation the browser adds by itself is kept.
Which browsers support speech to text
The browser decides whether a website may use speech recognition. Per MDN's browser compatibility data:
- Chrome has supported it since version 33 under the name webkitSpeechRecognition, and under the standard name since version 139. Edge follows Chrome.
- Safari has supported it since version 14.1 under the name webkitSpeechRecognition, on Mac, iPhone and iPad.
- Firefox offers it only in preview builds so far.
If your browser lacks it, the page says so and the Start button stays off. You can still type, paste, copy and download text in the box. The first time you press Start, the browser asks for access to the microphone. If you blocked it, allow the microphone for this site in the address bar and press Start again. To check that your microphone works at all, try the mic test.
Tips for clean dictation
- Pick the right variant. Choose the language and region you speak, such as English (India) or Portuguese (Brazil). The tool preselects your browser language.
- Speak in whole phrases. Talk at a steady pace and finish your sentence before you pause. Say the punctuation as you go instead of adding it later.
- Mind the room. Close the window, turn off the music and, if you have one, use a headset or a microphone close to your mouth.
- Fix as you go, or after. Correct a word in the box at any time. For tidy capitals afterwards, run the text through the case converter, and count it with the word counter.
- Hear it back. Paste your transcript into text to speech to listen to it, or keep longer drafts in the online notepad.
How to use the speech to text tool
- 1Pick the language you speak, press Start and allow the microphone when your browser asks.
- 2Talk. Gray words are first guesses; they turn into normal text when the browser is sure. Say period, comma or new paragraph to add punctuation.
- 3Press Stop, fix anything in the box, then copy the text or download it as a .txt file.
Frequently asked questions
Is this speech to text tool free?
Yes. There is no sign up, no time limit and no word limit. The tool uses the speech recognition built into your browser, so there is nothing to install. Freebiefy never receives your voice or your text.
Does my voice leave my device?
It can. Speech recognition is done by your browser, not by Freebiefy. Chrome sends the audio to Google's speech servers unless on-device recognition is used, and other browsers do not all document where the audio goes. Where your browser offers on-device recognition, tick Only on this device to keep the audio on your computer.
Why does the dictation stop after a pause?
Browsers end a recognition session on their own, for example after a stretch of silence. With Keep listening switched on, the tool starts a new session right away, so you can pause to think and carry on. Press Stop when you are done.
Which languages can I dictate in?
The list offers more than 70 languages and regional variants. Which of them your browser actually recognizes is up to the browser and its speech service. If one is not supported, the tool tells you so and you can pick another. Choose the variant you speak, such as English (United Kingdom) or Spanish (Mexico), for the best results.
Why does it not work in Firefox?
Firefox does not offer speech recognition to websites yet. According to MDN's browser compatibility data, it is available only in Firefox preview builds. Chrome, Edge and Safari support it, so open this page in one of them to dictate.
How do I add punctuation?
Say it. For English, say period, comma, question mark or exclamation mark, and say new line or new paragraph to break the text. The tool puts a capital letter at the start of each sentence. Untick Voice punctuation if you need to dictate the words themselves, such as "the period after the war".
Can I edit the text while I dictate?
Yes. The box is a normal text field. New words are always added at the end, so you can fix an earlier word while you keep talking. Copy the text or download it as a .txt file when you are done.