See what it hears
A small bar shows “Listening” and your words as they are recognized. It moves to the bottom when the field is near the top.
Free · No account · On-device
Saypen types what you say into the text field you’re using, on the websites you use every day. Chrome recognizes your speech right on your computer, so your voice doesn’t travel to a server and back.
How it works
Three steps on Saypen’s setup page, which opens by itself after you install it.
Click Allow microphone and choose “Allow while visiting the site”. Chrome asks once, on Saypen’s own page, not on every website.
Pick your language and click Download. Chrome fetches its speech model from Google: about 200 MB, once per language, usually in less than a minute.
Click into any text field, press Alt+Shift+D or the Saypen button in the toolbar, and talk. Press Esc when you’re done.
The setup page
The same page shows whether the microphone is allowed and your language is ready, and lets you try voice typing before you use it on a website.
To open it again later, right-click the Saypen icon in the toolbar and choose Options.
While you speak
A small bar shows “Listening” and your words as they are recognized. It moves to the bottom when the field is near the top.
Each finished phrase is typed with punctuation and capital letters. It’s a setting, on by default.
Saypen adds the space between phrases, and none in Japanese, Chinese or Thai, which don’t use them.
Press Esc, click Stop or say “stop listening”. It also stops after a silence (1 minute by default), when you switch to another tab or window, or when the page navigates or closes.
Voice commands
Format and fix your text without touching the keyboard.
Voice commands work in English, German, French, Spanish, Italian and Portuguese, and the setup page shows the words for your language. You can turn them off there.
Private by design
Saypen uses Chrome’s built-in speech recognition with on-device processing switched on. The audio goes from your microphone straight to Chrome’s recognizer on this computer. Saypen never records, saves or sends it, and Saypen itself makes no network requests at all.
The one download. The first time you use a language, you click Download and Chrome, not Saypen, fetches that language’s speech model from Google. Chrome keeps it for all websites, and from then on recognition works without an internet connection.
Read the privacy policy →Where it works
Ordinary text boxes, and the rich editors behind many mail, chat, AI-chat and social sites.
Text inputs, text areas, plain editable fields, and editors built on:
All 12 kinds were tested on our own test pages. It finds the field you’re in, also inside frames, and never types into password fields.
chrome://settingsChrome doesn’t let extensions run there. Saypen’s toolbar button shows a short explanation instead.
No text field selected? Your words stay in the listening bar with Copy and Close buttons, so nothing you said is lost.
Good to know
On a desktop computer, with on-device speech recognition available in that Chrome. The setup page tells you if it isn’t.
Built in or on a headset. Saypen uses it only while voice typing is on.
The ones Chrome can recognize on your computer, listed on the setup page. On Chrome 156 for Windows: about 16 languages, such as English, Spanish, French, German and Japanese. Albanian, Dutch and Ukrainian aren’t available yet.
Recognition is Chrome’s on-device model. It’s good for dictation, but names and specialist words can come out wrong.
Saypen is not a meeting transcriber. It doesn’t transcribe tab audio or other people’s voices; it listens to your microphone while you have it on.
No paid tier, no account, no ads and no minute limits.
FAQ
Yes. There is no paid tier, no account, no ads and no minute limits.
No. The microphone is open only while voice typing is on, or while the “Try it here” box on the setup page is listening. While it listens, Saypen’s toolbar icon shows a red “ON” badge. Your computer may also show that the microphone is in use. Voice typing stops when you press Esc, after a silence, or when you switch to another tab or window.
Yes, once your language is ready. The first time, Chrome downloads that language’s speech model from Google (about 200 MB). After that, recognition works without an internet connection.
No. In chat apps, “new line” makes a line break and doesn’t send the message. Send it yourself when you’re ready.
Saypen is made for Chrome on desktop computers. We have tested it on Windows so far. If on-device speech recognition isn’t available in your Chrome, the setup page tells you.
More answers, and fixes for common problems, on the support page.
Free for Chrome 139 or newer on your computer. No account, no minute limits, and your voice is recognized on your computer.
Get it on Chrome Web StoreMade by Iliria Studio