Free Speech to Text Dictation

Worked example

Set "Language" to English, leave "Voice commands and auto-capitalise" ticked, press the microphone button and say: "hello everyone comma welcome to the meeting period". When the recognizer returns those words, the tool turns "comma" and "period" into punctuation and capitalizes the start, so the box shows "Hello everyone, welcome to the meeting." Then say "new paragraph the agenda has three items" and it adds a blank line and "The agenda has three items". The header now reads "Transcript (67 chars) · 11 words", followed by the recording time, and "Copy" puts exactly that text on your clipboard.

Which settings to use, and the limits

• Choose the language before you start. The "Language" list has 51 languages and regional variants, including regional variants of English (US, UK, Australia, India), Spanish (Spain, Mexico, United States) and Portuguese (Brazil, Portugal). It starts on the language of your browser or of the page, falling back to US English, and it is locked while you are recording. If your browser cannot recognize the chosen language you get a message saying so; pick another one.
• Say punctuation out loud. With "Voice commands and auto-capitalise" ticked, which is the default and is remembered in this browser, English understands: period or full stop, comma, question mark, exclamation mark or exclamation point, colon, semicolon, new line or newline, and new paragraph. Spoken punctuation also works in Spanish, French, German, Portuguese, Italian, Turkish, Russian, Dutch, Polish, Arabic and Persian. In other languages, including Chinese, Japanese and Korean, spoken punctuation words are not converted, and only the sentence capitals are added where the script has them.
• Click in the text to dictate into the middle. The transcript box is editable. New speech is inserted where your cursor is, or at the end if you have not clicked into the text. Fix mistakes by typing. Once you edit by hand, SRT and VTT captions are rebuilt from your version, with the timing spread evenly across the recording.
• Pick the export you need. "Copy" copies the plain text, and "Share" appears where your browser supports it. "Download .txt", "Download .docx" and "Download .md" save transcript.txt, transcript.docx and transcript.md. Once speech has been recognized, SRT and VTT buttons save captions with timing. Tick "Add [0:12] timestamps to .txt, .md and .docx" to put the start time in front of each segment in those downloads.
• Your text survives a refresh. The transcript is saved in this browser as you go. After a refresh or a closed tab, "Restore last transcript" brings it back, and "Discard it" removes it. "Clear" empties the box and deletes the saved copy. The page also warns you before you leave with text you have not copied or saved. The screen is kept awake while you record where the browser allows it.
• Know where your voice goes. This tool uses your browser's own recognition engine, and it needs an internet connection. Chrome sends the audio to Google's speech service, Edge to Microsoft's and Safari to Apple's. Nothing is uploaded to LibraTool, and your transcript stays in this browser. For a recording that must stay on your device, use "Transcribe an audio or video file on your device" instead.

Common problems and how to fix them

It says "The browser's speech-recognition service couldn't be reached. Some browsers (such as Brave) block it — Chrome, Edge and Safari work."

Recognition happens in a service run by the browser's maker, so it needs a working internet connection. Check your connection and press "Try again". If you use Brave or a browser with its own privacy blocking, try Chrome, Edge or Safari. If nothing works, use "Transcribe an audio or video file on your device" with a recording.

It says "Microphone access is blocked. Allow the microphone for this site (the icon in the address bar), then press the mic button again."

The browser has no permission to use the microphone for this site. Allow it from the icon in the address bar, then press "Try again" or the mic button. On Android Chrome the message points to the menu (⋮), then Site settings, then Microphone.

It says "No microphone was found. Connect or enable one and try again."

The browser could not find a working input device. Plug in or enable a microphone, then press "Try again".

It says "Your browser does not support the Web Speech Recognition API. Works in Chrome, Edge, and Safari (desktop and mobile)."

Your browser does not offer the speech recognition feature this tool uses. Open the page in Chrome, Edge or Safari, or press "Transcribe an audio or video file on your device", which does not depend on the browser's speech service.

It says "Nothing heard for 10 seconds. Check that the microphone isn't muted and is the one selected in your browser."

No speech reached the recognizer for 10 seconds. Check that the microphone is not muted, that the right one is selected in the browser's own microphone setting, and that you are speaking in the language chosen under "Language". The message disappears as soon as speech is heard.

How it compares

Live recognition and file transcription suit different jobs. This page gives you text while you talk, which you can edit and export as captions, but it needs an internet connection, relies on your browser's service, has no speaker labels, and writes punctuation only when you say it. "Audio to Text" on this site transcribes a recording file in the page itself, so it also works without the browser's speech service and keeps the audio on your device, but it is not live. If you only need text typed into another app, your device's built-in dictation does that directly.