It builds up over time
Every word you look up and every passage you work through is saved. Next time you meet it, the answer is already there.
A reading companion for macOS
In any app on your Mac, select a passage and press ⌘⌥E — explain, translate, or rewrite it on the spot. What you look up stays with you.
macOS trial · 32MB · pick your AI model on first launch · Apple silicon
First time? Read the guide →
Every word you look up and every passage you work through is saved. Next time you meet it, the answer is already there.
The AI runs locally — nothing is uploaded. No account, no cloud, works offline.
Word, your browser, a PDF, a chat window — just select. Not happy with a result? Keep talking to it, or say what you want out loud.
A word, a concept, or a whole paragraph you cannot parse — explained in plain language.
Translate between languages, reading the surrounding text you did not select. The same word lands differently depending on whether the passage is about finance or a hobby.
Official notices, emails, posts — made clearer and on-tone, following the voice of the whole passage rather than standing out.
Text in a scanned PDF, an image, or a video that you cannot select — capture it, and explain, translate, or rewrite all the same.
It recalls what you looked up even when phrased differently, and over time speaks to the fields and level you tend to read at.
A key point from a book, a line worth keeping, a thread to research — select or capture it into your collection. When enough has piled up, “Export” lets the AI sort it, summarize each part, and hand you a tidy document.
The built-in on-device model handles everyday reading at no cost, with nothing uploaded. For sharper explanations or more natural translations, connect your own cloud key.
In any app, highlight the passage you want to understand or improve.
A small window appears right next to the text — no need to switch apps.
The result arrives at once. Everything you look up is saved, ready in the menu bar.
Notarized by Apple — opens cleanly.
Everything runs locally. What goes into Jinnang stays on your own machine.