Features
Live translated captions for what you hear, in-place translation for what you read — pick the surface that fits the moment. All of it runs in your browser.
What each surface does, when to reach for it, and what it needs — the short version. Every feature has a full page of its own.
An always-on-top window powered by Document Picture-in-Picture floats captions over slides, videos, and calls, movie-subtitle style. It goes wherever you drag it, at any size on any screen, and the captions reflow to fit while you're still presenting.
The browser extension overlays live translated captions right on the page you're viewing — YouTube, webinars, Google Slides. They stay put even over fullscreen presentation modes, and clicks outside the captions pass through, so nothing blocks your content. Sessions share the same account and history as the web console.
Stage mode turns a whole browser window into a caption screen: a large translation over a black backdrop with the source kept small — a type hierarchy legible from the back row. It relies only on the browser's fullscreen capability, so it also runs where the overlay window cannot, including Safari and Firefox.
Meeting interpretation splits the screen and runs both directions at once: you speak, they read; they speak, you read. No bots to invite and nothing to install — one screen share with 'Share audio' checked is the whole setup.
Voice to Voice speaks instead of only showing. Each sentence, once its translation has settled, is read aloud in a native voice through whatever your laptop plays to — the venue PA, a speaker, headphones — so the room can listen while the captions keep running. You pick the voice and hear a preview before anyone else does, and when speech runs ahead the voice catches up rather than drifting further behind.
A glossary pins product names and technical terms to your exact wording — the active set is injected automatically at session start, steering both recognition and translation. While you talk, smart context collects recurring terms and applies the ones you approve at the next sentence boundary, without dropping a caption.
Webpage translation lives in the extension: one shortcut translates the page you're reading right where it stands, with the translation under each line or replacing the text, and one more press restores the original exactly. One engine translates locally on your device, so the page text never leaves your browser; the other tunes the translation to the page's context and your glossary.
When a saved session ends, AI notes turn the transcript into a finished document — a one-page summary, meeting minutes, an earnings-call brief, or a recap of what you watched. Notes are written in your interface language whatever language was spoken, stick to what was actually said, and stay editable right in the record.
Whichever caption surface you pick, the pipeline underneath is the same.
Pick your translation from a catalog of 60 languages. Billing counts the time captions are on, never the number of languages — and running several languages at once is on the way.
Your browser connects straight to the speech engine with no relay hop, so captions follow the speaker as the words come.
The active glossary set rides along everywhere — overlay, stage, meeting interpretation, and extension alike; smart context joins in on console sessions.
Everything runs in your browser — three seconds from link to live captions. Want captions on the page itself? Add the browser extension. Prefer a menu-bar app? The macOS app is ready too.
Text size, line count, display mode, even the surface itself: reopen the overlay window or move to the stage mode without ending the session.
Event presentations, global meetings, content watching, and more — walkthroughs from preparation to the day of.
Whoever you talk to, wherever you are, connect with people everywhere in your own language.
