Turning Your Phone Into a Document Scanner: What Actually Gets Uploaded
Published 5 August 2026 · 5 min read
A landlord wants a photo of your lease, a clinic wants your insurance card, a form needs a "scanned copy" of your ID — and instead of digging out an actual scanner, you open a document-scanner app or website, point your camera, and it detects the edges, corrects the angle, and spits out a clean, flattened PDF. Convenient, and also a live camera feed of whatever identity document or contract you're holding, running through someone else's image-processing code.
Why scanning is a bigger ask than a regular photo
A normal photo you take and decide whether to share afterward. A scanner tool is built for the opposite flow — you point, it captures, and the processed result is often what you see, with the raw frames processed somewhere in between. If that processing happens on a server, every frame the camera captures while you're aligning a document — including the ones before you've decided the shot looks good — can pass through before you've made any choice about sharing it at all.
What people actually scan
The everyday use cases for a phone scanner skew toward exactly the documents you'd otherwise be careful with: government IDs, passports, insurance cards, signed contracts, medical forms, bank statements for a loan application. None of these are edge cases — they're the entire reason "scan a document" is a common enough need to have a dedicated app category. Which also means a scanner tool has a higher-than-average chance of touching something genuinely sensitive on any given use.
What "processed on a server" looks like for a scanner
Some scanner apps and sites work by uploading each captured frame (or the final image) to a server, running edge detection and perspective correction there with more processing power than a phone has, then sending back the cleaned-up result. For a receipt you're about to discard anyway, that's low-stakes. For an ID photo or a signed lease, it means the image exists on infrastructure you don't control for however long that service's retention policy allows, before you've even decided whether you're keeping the scan.
How FormatDog's Document Scanner avoids this entirely
FormatDog's Document Scanner runs edge detection, skew correction, and color-mode adjustment using JavaScript and canvas processing directly in your browser, reading frames straight from your camera via the browser's own camera API. The image never leaves your device to get cleaned up — the entire pipeline, from live preview to the final compiled PDF, runs locally. The companion GPS Camera tool, for stamping a photo with location and timestamp for field records, works the same way: the stamp is drawn onto the image locally, with nothing captured ever transmitted to process it.
You don't have to take that on faith
Open developer tools (F12, or right-click and choose "Inspect"), click the Network tab, then scan a document and watch what happens while you capture and process it. If frames or the final image are being uploaded, you'll see outgoing requests carrying that data appear in the list. If nothing shows up while the scan completes and the PDF gets built, it's happening locally, exactly as described.
The honest caveat
Running entirely in the browser does mean scan quality depends on your own phone's camera and processing power rather than a server's more capable image pipeline — for most documents under normal lighting, that difference isn't noticeable, but a server-side tool with more sophisticated correction might occasionally handle a genuinely difficult angle or lighting condition better. For anything involving an ID, a contract, or a medical form, that's a trade most people would make gladly.