Frequently asked questions
Everything people usually ask before trusting a document scanner with their paperwork.
Is ScanPDF free?
Yes, and unconditionally so. There is no account, no sign-up, no page limit, no watermark, no paid tier and no trial. The whole thing is MIT-licensed open source, and there is no server to charge you for.
Are my documents uploaded anywhere?
No. Image processing and PDF assembly both run in your browser, and the app has no backend to send anything to. Your photos are read from disk by the page, held in memory, and written back out as a PDF download. See the privacy page for how that is enforced rather than merely promised.
Which image formats can I use?
Anything your browser can decode as an image: JPEG and HEIC from a phone camera, PNG, WebP, AVIF, GIF, BMP. Photos straight off an iPhone or Android device work without conversion, and their EXIF orientation is handled automatically.
Is there a page limit?
No fixed one. Pages are exported sequentially rather than all at once, so memory use stays flat and long documents are fine. The practical ceiling is your device's RAM with very large source photos, not a rule in the app.
Does it work on iPhone and Android?
Yes, in any current mobile browser. On a phone the layout switches to a single panel with an editor and preview toggle, and a "Take photo" button opens the rear camera directly, so you can photograph and scan a document without saving it to your camera roll first.
Does it work offline?
Yes. After your first visit the browser keeps a copy of the app on your device, so it opens without a network. The scan engine, a download of about 10 MB, is kept as well once you have scanned a page or installed the app, and from then on everything works offline. If you want a genuinely airgapped setup, run the container described below - it serves the whole app, including the OpenCV build, from your own machine.
You can also install ScanPDF to your home screen. On Android it then appears in the share sheet: select photos in the gallery, tap Share and pick ScanPDF. On iPhone and iPad, use Add to Home Screen in Safari's share menu; iOS does not let web apps receive shared photos, so add them from inside the app there.
How good is the output?
Export warps the full-resolution original, not the downscaled copy used for detection, capped at 3500 px on the long side and written at roughly 220 DPI on an A4 page. That is enough to print or to run OCR over. The black and white filter uses background division rather than a hard threshold, so shadows flatten out and text stays anti-aliased instead of turning jagged.
Can I reorder or delete pages?
Yes. Drag the page thumbnails to reorder them and delete the ones you do not need; the order there is the order in the exported PDF. Rotation, corner adjustments and the filter are all per-page, so one bad photo in a batch never forces you to redo the rest.
Why did automatic corner detection miss?
Almost always because there was no edge to find: white paper on a white table, a corner outside the frame, or a hard shadow across the page reading as a false edge. Drag the corners in manually and everything downstream works identically - detection is a convenience, not a requirement. The scanning guide covers how to avoid it.
Can I run it on my own machine?
Yes. A prebuilt container image is published to GitHub Container Registry, and each release also ships a plain zip of static files for any web server. Self-hosting additionally gives you the strict Content-Security-Policy that makes the browser itself block outbound requests. Instructions are in the README.