Scan documents to PDF with your phone camera

Tap Start camera and the browser asks for permission. The tool requests the rear camera first and asks for a frame of about 2048 by 1536 pixels; the phone supplies the closest size it has. Each tap of Capture page grabs the current video frame and adds it to the file list as a JPEG named scan-01.jpg, scan-02.jpg and so on. Decline the permission and a short note replaces the preview; the file picker still works.

The Scan look checkbox is on by default, hinted as Grey, high contrast. It drops the color, raises the contrast and lifts the brightness a little, so print reads darker and paper reads whiter. That is the entire effect: no edge detection, no automatic cropping, no perspective correction, no shadow removal. The live preview carries the same filter, so what you see before you tap is what gets saved. Switch camera flips to the front camera and back.

Because nothing is corrected afterward, the shot itself has to be right. Put the page on a plain surface in even light, ideally from a window to the side rather than a lamp overhead, so the phone does not shadow the text. Hold the phone flat over the page, parallel to it, and fill the frame; a tilted phone produces a keystone-shaped page that stays that way. Keep still for a moment before you tap, because the capture takes whatever the video shows at that instant.

Turn Scan look off where color matters: a photo, a colored chart, a form with red stamps. The added brightness can also wash out faint pencil or light grey print, so toggle the checkbox and watch the preview before capturing. Afterward the file list shows only names and sizes, not the pictures, so a bad shot is easiest to catch as you take it. To drop one, tap the X on its row; rows are named in capture order, so note which number went wrong.

Match each photo makes every PDF page exactly the size of its image, with no margin. It measures the page in points, one per pixel, so a 2048 by 1536 capture becomes a page around 28 by 21 inches: fine on a screen, awkward at a printer. Pick it for odd shapes such as till receipts.

A4, Letter, Legal, A3 and A5 give a fixed page with a quarter-inch margin, the photo scaled to fit inside it and centered, so a scan of an A4 sheet prints without being cut off. Orientation on Auto turns the page sideways whenever a photo is wider than it is tall, which suits a mix of portrait pages and a landscape spreadsheet. Set Portrait or Landscape only when Auto guesses wrong. The setting disappears under Match each photo, where there is no page to rotate. A camera frame is roughly four by three and no paper size is, so a fixed page always leaves extra white on two sides.

The file list shows the captures in the order you took them; drag a row, or use the up and down arrows, to reorder. You can mix in photos already on the phone with Add more files, which takes JPG and PNG. That picker allows 20 files at a time, the ceiling for gallery photos. Live captures are not counted against it, but every page sits in memory until you build, so on an older phone build in parts and join them with Merge PDF.

Type a name in Output file name, then tap Build PDF from N pages, where N is the number of rows. The panel that follows has a Download button carrying that name. Everything is held in the tab, so a reload, including one the browser does on its own while you are away, wipes the captures and you start again.

The output is a PDF of pictures, so text in it cannot be selected or searched. Run the file through OCR PDF to add a hidden text layer under the image; recognition runs in the browser with Tesseract, covers English only, and does best on clean printed text rather than handwriting. If the file is large, Compress PDF re-encodes the JPEG pages at a lower quality. Sign PDF places a picture of a signature on a page, and Protect PDF encrypts the file with a password.