VGrabThe short answer
In the extension it reads and never sends. It inspects network requests and the current page in your browser to work out whether there is a saveable video; detection, downloading, merging and writing to disk all happen locally. It does not upload which pages you opened or what you downloaded, carries no analytics script and does no error reporting. Uninstalling clears everything — it writes nowhere else on your computer.
So what does it read
To find a video it looks at the current page's network requests (where the playlist is) and the page's public tags (title, cover, duration). It only looks; nothing leaves: this is used to list candidates and pick a quality in the popup, and that is the end of it.
When fetching a video it carries the login state you already have on that site (otherwise members-only content can't be retrieved), but it never reads, stores or changes your credentials — it lets the browser decide, exactly as when you press play yourself. And only for servers related to that page: the page itself, where the playlist lives, and the hosts the player already talked to; no other server gets it.
Four things, all local
- Detection records: which addresses this page produced, in browser memory only, cleared when the tab closes.
- Pieces of a running download: held in the browser's own storage, cleared once moved to your download folder; an abandoned download is cleared at once.
- Download history: where, how big, when — in this browser only, with one-click clearing on the manage page.
- Settings: quality, parallel lanes and so on — local only.
What about the website
The website (the one you're reading) is separate: like most sites it uses Google Analytics and Google AdSense with cookies, and European visitors see a consent box first. That is only on the website pages — the extension loads none of the website's scripts.
The extension opens website pages in two cases: the downloads page (progress and history in a full tab, roomier than the popup), which carries no ads and uses no analytics cookie about your downloads; and the feedback page when you click "Having trouble?" in the popup. What you type there is emailed to us only when you press Send; whether the report includes the full address of the page you were on is a switch on the form.
Check for yourself
The extension is plain front-end code shipped in the package with no remote loading. To verify: open your browser's extensions page to see the permissions it requests; open the network panel in devtools, download a video, and watch it connect only to the video's server, never to our domain; and the ZIP package is the source — unzip it and read.
How this relates to the privacy policy
This page is the plain-language version; the formal one is the privacy policy, which lists what the extension collects (nothing), what the website collects (analytics and ad cookies) and what the feedback page collects (what you type). If the two ever differ the policy wins; tell us on the feedback page and we'll fix it.
Questions
Why the "read network requests" permission? Web video is usually a stream whose address isn't in the page, only in the player's network requests; without seeing them the video can't be found. It reads the address, not the contents.
Why all-sites access? VGrab recognises video formats, not specific sites, so it must work anywhere; a narrow domain list couldn't mean "download from any site".
Why the "modify request headers" permission? Many video sites block hotlinking and only accept requests that come from the page. When the extension fetches on the page's behalf it adds the page's real referer so the server lets it through — the real page address, not a forged identity.
Could an update quietly start collecting data? Every version's changes are listed in the release notes on the site; any privacy policy change is announced there too.