How to protect your privacy when you use online file tools

A page can load from the internet and still do the file work on your computer. That is what QuickConvert’s tools do. PDF Merge, PDF Split, PDF Compress, PDF to Images, Image Compress, Image Convert, Image Resize, and HEIC to JPG read the file with the browser’s file APIs, process it in the tab, and download the result from memory. The word counter, Hijri converter, QR generator, and Arabic amount tool do the same with text you type: it is not posted anywhere. The PDF library is pdf-lib, and PDF pages are drawn with PDF.js when you ask for an image of each page. Images are drawn with canvas, and HEIC photos are decoded in the page. A batch of images, or a PDF split into many pages, is zipped with JSZip in the page. There is no request that posts the file bytes to a QuickConvert server, because there is no processing server in this design.

That is a stronger privacy property than “we delete your file after an hour,” which still means the file was sent. It is not the same as “nothing about your visit leaves the device.” The page itself was downloaded from the host. The host can log that a browser asked for the page. Optional measurement and ads can do more. The useful habit is to separate those two facts and decide whether each one is acceptable for the document in front of you.

When you drop a file, the browser keeps it in memory for the page. PDF tools check the type, refuse anything over 80 MB, and refuse encrypted files rather than asking you to type a password into the site. Image compress, convert, and resize refuse anything that is not JPG, PNG, or WebP, and anything over 40 MB. HEIC photos use the separate HEIC to JPG tool, which has the same 40 MB cap. Object URLs used to decode images are revoked when that image is finished, including when processing throws. The download is started from a temporary object URL that the page revokes about a second later. Close the tab and those working copies go with it. There is no account and no library of past files to delete later.

You can confirm the shape of this, without special tools, by loading a tool page and then turning the network off before you drop a file. The merge, split, compress, convert, or resize can still finish, because it is not waiting on an upload. The first load of the page does need the network. A later visit needs it again if the browser did not keep the scripts.

This also means QuickConvert cannot recover a file for you, and cannot “finish the job” if your tab crashes. Keep your own original until the download is open and correct.

Hosting the site on Vercel produces ordinary server logs: an IP address, the browser type, the URL that was requested, and a time. Those logs are about the request for the page, not about the contents of a PDF that was never in the request.

The public site loads Google Analytics 4 with measurement ID G-T7YG5RCWKQ. Analytics receives things like which pages were opened, a coarse location, and the kind of browser and device. It sets its own cookies, commonly including a cookie named _ga, to tell browsers apart. It does not receive the file you dropped. You can limit this with your browser’s cookie controls or with Google’s Analytics opt-out add-on. The Privacy Policy states the same measurement ID.

The site is also an AdSense site, publisher ID ca-pub-5685826130857195, which is the ID in ads.txt and in the site’s account meta tag. When an ad unit is filled, Google and its advertising partners may set cookies and use them to measure and to choose ads, including ads based on activity on other sites. Ad personalization can be adjusted at Google’s ad settings. Ads on QuickConvert are placed in the articles and guides, not on the empty tool, the progress line, or the download status, and not on the 404 page. An ad request still does not include your PDF or image. If you would rather not load advertising cookies, a browser that blocks third-party cookies or an ad blocker will stop the ad. The tool still runs.

The theme control stores a light, dark, or system preference in the browser’s local storage so the next visit matches. That value stays on the device. It is not an account.

Use a tool whose processing you understand for anything you would not want copied to a stranger’s server: identity documents, medical papers, unpublished work, client files. “In the browser” on this site means the tools linked above, including the text tools that never take a file. If you follow a link to a different site, read that site’s own description. A progress bar that says “uploading” means the file is leaving.

Do not type a PDF password into a web form to “unlock” a file. This site will not ask. A site that asks is asking you to hand over both the file and the secret.

Prefer a copy, not your only original. Local processing can fail when memory runs out, and a failed run should leave the source untouched. Check the download’s page count or the image’s appearance before you delete anything.

If the document is so sensitive that even a page-view log feels like too much, do not open it on a website at all. Use an offline application. Privacy on a public site is a set of limits you can verify, not a claim that the internet was not involved in loading the page.

Questions about this policy can go to the address on the contact page. The full cookie and analytics description is the Privacy Policy.