Speech to Text (Voice Typing)

About the Speech to Text (Voice Typing)

Speak and watch your words appear: choose a language — Indian English, Hindi, Marathi, Tamil, Bengali, Telugu, Gujarati and more — press Start, and dictate. Say comma, full stop, question mark or new line to punctuate (in Hindi too: अल्पविराम, पूर्ण विराम). The text lands in an editor where you can correct it at any time, and you can copy it or download it as a .txt file.

Recognition is done by your browser’s speech engine (the Web Speech API) in Chrome, Edge and Safari. Where the browser can recognise your language on this device — desktop Chrome can, after a one-time download of the language — the page uses that, and your voice never leaves the device. Otherwise the page tells you before you start that the browser sends the audio to its speech service.

Common uses

• Writing emails, messages, notes and first drafts faster than typing, especially on a phone.
• Typing in Hindi or another Indian language without learning a keyboard layout.
• Taking notes hands-free, or writing when typing is painful or difficult.

Spoken commands

• English: comma , · full stop or period . · question mark ? · exclamation mark ! · colon : · semicolon ; · open bracket / close bracket ( ) · open quote / close quote “ ” · hyphen - · dash — · new line · new paragraph · delete that (removes the last phrase).
• Hindi: पूर्ण विराम । · अल्पविराम , · प्रश्न चिह्न ? · नई लाइन · नया पैराग्राफ · इसे मिटाओ (removes the last phrase).

Period, colon and dash are also everyday words, so they count as punctuation only at the end of a phrase — pause briefly before saying them. Turn Spoken punctuation off if you want every word typed as said.

On this device or online?

Browsers recognise speech in one of two ways. On-device: the language model runs on your computer, nothing is sent anywhere, and it works offline. Desktop Chrome offers this for many languages after a one-time download, which this page can start for you; Chrome rates these on-device models for short phrases, so long dictation can be less accurate than online. Online: the browser sends the audio to its maker’s speech service — Google for Chrome, Microsoft for Edge, Apple for Safari — and gets the text back. Before you press Start, the page shows which one applies to your language and browser; tick Only on this device to make sure audio is never sent.

Which browsers work

Chrome (desktop and Android), Edge on the desktop and Safari (Mac, iPhone and iPad; all iPhone and iPad browsers use Safari’s engine) support voice typing. Firefox does not offer speech recognition to websites yet, except in test builds. On iPhone and iPad, Siri or Dictation must be turned on in Settings.

Tips for better results

• Use a headset or the phone close to your mouth, in a quiet room.
• Speak in phrases of a few words and pause briefly between them.
• Choose your language variety (English (India), not English (United States), for Indian names and places).
• Say punctuation as separate words, with a short pause before period or colon.

Limitations

• Accuracy depends on the browser’s speech engine, the microphone and background noise.
• This page types what you say live; it cannot transcribe recorded audio files.
• Browsers end a listening session after silence or about a minute; the page starts a new one at once, which on Android can play a short sound.
• Spoken commands work in English and Hindi; other languages are typed as recognised.
• Firefox is not supported, and the language list depends on the browser’s speech service.

Privacy

Unless on-device recognition is available and in use (the page tells you), your browser sends your audio to its own speech service — Google in Chrome, Microsoft in Edge, Apple in Safari — to turn it into text. MySmartCoPilot never receives your audio or your text.