Speech to Text

What this tool does

Speech to Text turns what you say into written text in real time using your browser’s speech recognition. Choose the spoken language, see words appear as you talk, say punctuation such as “comma” and “new paragraph” in English, add timestamps, then edit, copy or download the transcript.

Most people speak about three times faster than they type, around 130 to 150 words per minute against 40. Dictation is ideal for first drafts, long messages, notes after a meeting, or for anyone who finds typing painful.

This tool uses the Web Speech API’s SpeechRecognition interface. Your browser streams the microphone audio to its recognition service, which returns interim guesses while you talk and a final transcript for each phrase. Browsers end a session after a pause or about a minute, so the tool restarts it automatically until you press Stop. Be aware of where the audio goes: Chrome sends it to Google, Edge to Microsoft and Safari to Apple or its on-device recogniser. The transcript itself lives only in this page.

How to use Speech to Text

1. Choose your Spoken language and accent, for example English (United Kingdom) or Hindi.
2. Decide whether to show words while speaking, use spoken punctuation and add timestamps.
3. Click Start listening and allow microphone access.
4. Speak in full phrases; say “comma”, “full stop”, “question mark” or “new paragraph” for punctuation in English.
5. Click Stop when you are done.
6. Correct any mistakes in the Transcript box, then Copy or Download it as a text file.

Tips for better results

• Pick the regional variant you actually speak; en-IN and en-GB models handle those accents noticeably better than en-US.
• Use a headset or sit close to the microphone in a quiet room. Background music and echoing rooms cause most errors.
• Speak in complete phrases rather than word by word; the recogniser uses context to choose between similar-sounding words.
• Turn on timestamps when transcribing a talk so you can find passages again in the recording.

Common use cases

• Drafting emails, essays or blog posts by voice.
• Capturing meeting notes or ideas while walking.
• Making typing easier for people with RSI or limited mobility.
• Practising pronunciation in a foreign language and checking what the recogniser hears.

Common mistakes to avoid

• Using Firefox, which does not provide speech recognition to websites.
• Dictating confidential information without considering that the audio is processed by the browser vendor’s service.
• Expecting it to transcribe a recording file: it listens to the live microphone only.