How to scan a document with your phone's camera
Take a photo of each page, drop the photos on this page and press Save PDF. Each page is found, flattened and cleaned on its own; check the pages, then save one PDF, searchable if you like.
Add the photos
One photo per page, or one of an open book. Pages appear in the order the photos were taken. Add more at any time.
Check the pages
Drag the pages into order and delete the ones you do not want. If a page came out cut or skewed, choose Edges and drag the corners, or the handles in the middle of each edge for a page that bows.
Pick a look
Colour keeps stamps, highlighting and photos; grey is lighter on the eye; black and white is the smallest and sharpest for text.
Save
A PDF on A4, Letter or each page's own size, with searchable text if you like, or every page as a JPG or PNG.
Curved book pages, made flat
A page that bows up from a book's spine bends every line of text. The scanner finds each line with PaddleOCR's text detector, fits the curve of the paper to all of them at once, and draws the page again with the lines straight; a second, finer pass takes out what the first left. Lines in a table or a heading set at an angle are left out of the fit, and a flat page is left as it is.
A photo of an open book is split at the fold, where the paper darkens towards the spine, into a left and a right page, each straightened on its own. Turn Two pages off to keep a spread whole.
Shadows and looks
The paper's colour is measured everywhere on the page, from the light parts between the letters, and each pixel is divided by it: shadows, a lamp's warm light and the shading near a fold all come out white, and the ink stays dark.
| Look | Best for |
|---|---|
| Colour | Forms with stamps and signatures, highlighted text, pages with photos |
| Grey | Printed text and pencil, a softer look than black and white |
| Black and white | Text only: the sharpest and smallest files (Sauvola's threshold, after the shadows are gone) |
Taking a good photo
- Lay the page on a darker surface so its edges stand out
- Hold the phone above the page, not at a steep angle, with the whole page in the picture
- Daylight beats a lamp; keep your own shadow off the page
- For a book, press it open gently: a gentle curve is straightened, a deep fold hides text
Where do the photos go?
Nowhere: they stay on your device. The pages are found, straightened and cleaned in your browser, and the PDF is written there too. The line finder (19 MB) and the text reader (3.5 MB plus a language) download from this site once and are kept in your browser for next time.
Specs and browser support
| Input | JPG, PNG, WebP, AVIF, GIF, BMP, several at once; photos over 24 megapixels are scaled down |
|---|---|
| Output | PDF (A4, Letter or each page's size), searchable or not; JPG or PNG per page, several in a ZIP |
| Pages | Up to 3508 pixels on the long side: A4 at 300 dpi |
| Engine | Page finding, flattening, straightening and cleaning written for webdave; PaddleOCR PP-OCRv5 text detection (Apache-2.0) on onnxruntime-web (MIT); Tesseract 5 (Apache-2.0) for searchable text; pdf-lib (MIT) writes the PDF |
| Download | Line finder 19 MB with its runtime, on the first page straightened; text reader 3.5 MB plus 0.3 to 3.8 MB per language, when searchable text is on; once, then kept on your device |
| Browsers | Current Chrome, Edge, Firefox and Safari, on computers and phones |
| Price | Free, no account, no watermark |
Questions
No. Everything runs in your browser: finding the page, straightening, cleaning, reading the text and writing the PDF. The line finder and the text reader download from this site once and stay on your device.
Yes. It finds every line of text on the page, fits the curve of the paper to them, and draws the page again with the lines straight. A photo of an open book is split into its two pages first.
Put the page on a darker surface, hold the phone above it rather than at a steep angle, and avoid your own shadow. Daylight is best; shadows that remain are taken out.
Turn on Searchable text before saving. The text is read in the language you pick and laid over each page as invisible text, so you can search, select and copy it.