"Your files never leave your device" is on every browser-tool landing page now. I build one of these sites, so here is what that claim really covers — and where it quietly stops.
1. "No upload" ≠ "offline." The file is read with the File API inside the tab, but the page still loads its PDF and image libraries from a CDN. Open the tool with no network and it won't start. I wrote up exactly what happens step by step in what the browser does with your file.
2. Client-side PDF compression usually means rasterising. The simplest way to shrink a PDF in JS is to draw each page to a canvas and re-encode it as JPEG. Great for scans. Terrible for a text PDF: it can get bigger, and the text is no longer selectable. Our Compress PDF does exactly this, so the page says so before you click.
3. A black box is not a redaction. Drawing a rectangle over text hides it on screen and in print, but the text layer underneath is still there — copy-paste or a text extractor will find it. That's why our Redact PDF is labelled "visual redaction" in plain words. If it's a legal document, flatten to an image or use a tool that removes the content stream.
4. Photos carry more than pixels. Before you post a picture, check the EXIF block — GPS coordinates are often in there. An EXIF viewer that can save a stripped copy takes a few seconds and runs entirely in the tab; a lossless strip keeps the pixels untouched.
Bonus: HEIC. Browsers don't decode HEIC natively, so in-browser converters ship a decoder (WASM). Expect a fixed quality setting and only the first frame of a Live Photo — details in what HEIC to JPG keeps and drops.
Takeaway for builders: state the limits on the tool page. Users trust a tool more when it tells them when not to use it.
Disclosure: I run PC Tools Online, the site linked above. Happy to answer implementation questions in the comments.
Top comments (0)