Privacy
Your files never leave your device. We count page visits and tool use, with no way to tell who you are, to learn how many people use wtfpdf and what to build next.
Last updated 9 October 2026
Your files
Every tool runs in your browser. The PDF engine is compiled to WebAssembly and works on your files in memory on this device. Files are never uploaded, and nothing you type, including passwords and watermark text, is sent anywhere.
The site's security policy only lets the page connect to wtfpdf.co and to Cloudflare Web Analytics, so the browser itself blocks any attempt to send a file somewhere else.
Make searchable downloads its text recognition engine, and a file for each language you choose, the first time you use them. They come only from wtfpdf.co, as plain requests for fixed files, and only after you press the button that shows their size. The requests carry nothing about your file. Our host can see which language files were downloaded, the way it sees any page load. After that, Make searchable works offline.
wtfpdf uses your camera only on the Scan to PDF and Images to PDF pages, and only after you tap Open camera or Take photos. The camera turns off when you tap Done, leave the page or switch to another tab, and nothing records video. Photos stay on your device like every other file.
Page visits
wtfpdf.co is hosted on Cloudflare Pages and uses Cloudflare Web Analytics to count visits. For each page view it records:
- the page address, without anything after a question mark
- the site that linked to it
- your country
- your browser, operating system and type of device
- how fast the page loaded
It sets no cookies. Cloudflare says it does not follow people across sites. Cloudflare keeps this data for six months.
Tool use
When a tool finishes, fails, or cannot open a file, the app sends one short record to wtfpdf.co. It holds:
- which tool it was, and whether it worked
- the kind of error, if there was one, such as "damaged file" or "unsupported feature"
- how many files and pages went in
- the total size, as a range such as "1 to 10 MB"
- how long it took, as a range such as "1 to 5 seconds"
- whether a file was password protected, and whether you were offline
- whether a file was signed or certified, but never who signed it, when, or anything from its certificates
- your browser and operating system, as a family such as "Firefox on Windows"
- the settings you picked from a list, such as the rotation angle, the image resolution, the watermark font, whether Compress kept quality or made a smaller file and at which setting, or roughly how much Compress saved
- for Crop, whether you trimmed margins or split pages in half, and for a split, which way you cut and the order of the halves, but never where you cut
- for Fill form, roughly how many fields the form has, how many you filled, how many could not be written and how many needed our font for letters such as Greek or Cyrillic, as ranges such as "6 to 20", and whether you flattened it, but never a field's name or what you typed in it
- for Split, whether you split by page ranges, every page or at blank pages, and at blank pages whether the file was scanned on both sides, but never which pages
- for Remove blank pages, whether you chose Strict, Normal or Lenient, whether you marked or unmarked any pages yourself, and roughly how many pages were removed, as a range such as "2 to 5", but never which pages
- for Pages per sheet, whether you made pages per sheet or a booklet, how many pages per sheet and in which order, the binding, which edge the printer flips on and whether you allowed for creep, the sheet size from the list, the orientation and scaling, and whether you added borders or crop marks, but never the margins, the gutter or the paper thickness
- for Images to PDF, the page size, orientation, margins, placement and image quality you picked from the list, whether you made one PDF or one per image, whether any image was animated and which frames you kept, the image format, or "mixed" when they differ, whether your browser decoded any image, whether the images came from files, the camera or both, whether you saved the result as a PDF or as images, and whether a file was not an image we can read or one your browser could not decode, but never a custom page size or anything from the images
- for Flatten, whether you flattened form fields only or comments too, and whether there was anything to flatten
- for Scan to PDF, the same as for Images to PDF, and whether the photos came from the camera, its preview or your camera app, whether you cropped none, some or all of them, whether auto capture took none, some or all of them, roughly how many you retook, as a range such as "2 to 5", and whether the camera was blocked or missing, but never a photo or anything in it
- for Clean up scan, whether all, most, some or none of the pages are scans, which steps were on, the filter you picked from the list, and roughly how much smaller the file got, but never a page or anything on it
- for Verify signatures, roughly how many signatures, timestamps and empty signature fields the file has, and how many signatures were intact, changed after signing, altered or could not be checked, as ranges such as "3 to 5", but never who signed, when, or anything from a certificate
- for Make searchable, which languages you chose, whether you chose every scanned page, some of them or pages that already had text, whether you also made the file smaller and at which setting, whether the run downloaded the text recognition files first, and roughly what share of pages had text found, as a range such as "51 to 99%", but never the text it found
The record never contains a file, a file name, page content, a password, anything you typed, or anything that identifies you or your device. We don't store your IP address with it, and there is no user ID, so two records can't be linked to the same person. The records are kept in Cloudflare Workers Analytics Engine for three months. We may keep totals built from them, such as the number of PDFs merged, for longer.
If you use wtfpdf offline, records wait in your browser's storage and are sent the next time you're online.
Turn off usage counts
This stops tool use records from this browser and deletes any that are waiting to be sent. Browsers that send Global Privacy Control are already opted out. Page visit counts are separate. Content blockers such as uBlock Origin, and browsers such as Brave, block those.
What stays on your device
- The app's files, cached by a service worker so wtfpdf works offline.
- The tools you've used, to show them under Recent. This list is never sent anywhere, and Settings clears it.
- Saved signatures, one signature and one set of initials, kept only when you tick Remember on this device. They are never sent anywhere, and anyone who uses this browser can use them. Settings deletes them.
- The text recognition files Make searchable downloads, so it works offline. Settings removes them.
- Your theme, your choice about usage counts, and any records waiting to be sent.
- Any feature switches set through the page address, which we use for testing.
Clearing this site's data in your browser removes all of it.
What we never do
- Upload, read or keep your files
- Use cookies, accounts or advertising trackers
- Sell or share data with anyone for advertising
Why we count
The counts tell us how many people use wtfpdf, which we share publicly as totals such as "PDFs merged this month". They also show which tools people rely on, which files the engine fails to open, and where tools are slow, so we know what to fix and what to build next.
Questions
Ask a question or raise a concern in WTFPDF/community.