How it works
One idea: looking something up should not cost you your place.
The idea
You select text in a PDF and a Find meaning action appears. Choosing it opens Google Search results in a panel inside the reader. Close the panel and the page is exactly where you left it, with no tab to find your way back from. The name is the same idea: DND stands for Do Not Disturb.
On a computer, Ctrl+E chooses Find meaning and Esc closes the panel. On Android, press and hold a word, drag the handles, then choose Find meaning; Back closes things one layer at a time, so it closes the panel before it leaves the document.
Technically, the selected text is put into a search address and the panel navigates to it when you click, like a browser. DNDReader never reads, scrapes or automates the results page.
What it doesn't promise
- Google decides what the panel shows. An AI Overview sometimes appears and often doesn't. DNDReader does not promise one.
- You can't sign in to Google in the panel. Google blocks sign-in inside embedded browsers, so results are signed out and not personalised.
- A consent screen can appear in the panel, and occasionally a CAPTCHA. You deal with it there, as you would in any browser.
- Looking something up needs an internet connection. Reading a PDF does not.
- Text-based PDFs only. A scanned page has no text to select, and there is no OCR.
These are limits of how the panel works, not bugs waiting for a fix. DNDReader shows Google Search; it is not an AI assistant and has none of its own.
One question isn't settled. Whether Google's terms permit loading Search inside another app's panel has not been established by this project. Google's terms are written around browsers, and nothing here concludes that any use is exempt. Google may also change how embedded browsers behave without notice.
Why the downloads differ in size
The search panel needs a browser engine. On Windows it uses Microsoft's WebView2, and on Android the phone's own system WebView, which is why those downloads are small. Linux has no built-in web view DNDReader can use, so the Ubuntu package carries its own Chromium, which is why it is large and why it takes updates from you rather than from Ubuntu.
What stays on your device
What leaves it: the text you select, as a Google search, when you choose Find meaning. Your PDF is never uploaded. There is no telemetry, no analytics, no crash reporting and no server of ours. Only your preferences and the paths of recently opened documents are stored locally.
Google's cookies are loaded in the panel as they would be in a browser. What happens to them differs by platform:
- Ubuntu
- The search panel's cookies and storage are held in memory only and discarded when DNDReader quits.
- Windows
- Windows can't keep the panel's data in memory, so it uses a temporary folder that is cleared and deleted when you quit. During a session that data is on disk, and after a crash a leftover folder can remain until the next launch removes it.
- Android
- Android can't keep the panel's data in memory either, so it lives in the app's private storage while the app runs, where other apps can't read it. That storage is deleted the next time the app starts. It is not gone the moment you swipe the app away.
On Android, a choice you make on a Google consent screen lasts until the app is closed and reopened. On every platform the consent screen can appear again the next time you start the app.
More detail
The full reasoning, including the options that were rejected, is in the project's implementation plan and the README. Found something that doesn't match this page? Open an issue.