Scan to PDF (corner correction + OCR)
Turn a quick phone photo of a document or a book page into something that looks like it came out of a real scanner. Upload a photo and the tool automatically finds the four corners of the page, then flattens the shape into a straight, rectangular image even if it was shot at an angle or from the side. Pick a Black & White document mode to make faded or unevenly lit text crisp and readable, or Sharper to boost contrast without going fully monochrome. When you're happy with the result, optionally turn on OCR (text recognition) to bake an invisible, selectable and searchable text layer into the final PDF. Every step — corner detection, perspective warping, filtering and OCR — runs locally in your browser. Nothing is ever uploaded to a server, which matters if you're scanning IDs, contracts, receipts or personal notes.
Loading the tool...
How to use it
- Tap 'Take a photo' or 'Choose files' to add one or several document/book page photos at once.
- Four green corner handles appear automatically on each photo. If the detected position is off, drag the dots onto the real page corners and press 'Apply to this page' to flatten it.
- Pick Original, Black & White document, or Sharper for the look you want, then reorder, rotate or delete pages in the page list below.
- Choose a page size (A4, Letter, or Original size) and image quality, turn on OCR if you want selectable text, then press 'Create PDF' to save it instantly.
Open-source used here
Everything runs client-side, with no server round trip. Corner (edge) detection does not rely on a third-party computer-vision library; instead it is a small custom JavaScript routine that converts the photo to grayscale, measures edge strength with a Sobel-style gradient, and then searches in four diagonal directions for the farthest strong-edge pixel to approximate each corner. This works well whenever the document has reasonable contrast against its background, and you can always fix it by dragging the handles. The perspective warp that flattens the quadrilateral into a straight rectangle is likewise a small amount of custom math (a homography solved with Gaussian elimination and bilinear-sampled rasterization) rather than an external library.
- pdf-lib (MIT license) — assembles the processed page images into the final PDF file.
- Tesseract.js (Apache-2.0 license) — the OCR engine that only runs when you turn on 'OCR', producing the invisible text layer inside the PDF. Korean/English recognition data is downloaded once and cached for next time.
⚠️ Copyright note: scanning an entire copyrighted work such as a book, magazine or newspaper and sharing or uploading the result can infringe copyright, even when the scan itself is technically easy to make. Please use this tool for personal, non-distributed purposes — contracts, receipts, homework, handwritten notes, forms you need to submit — rather than producing and distributing full scans of books you don't own the rights to.
Frequently asked questions
What if the automatic corner detection gets it wrong?
You'll see four green dots on the photo. If the detected position is off, just drag each dot onto the real corner of the page, then press 'Apply to this page' to warp it correctly.
Is it okay to scan a whole book?
It's technically possible, but scanning an entire copyrighted book and sharing or uploading it can infringe copyright. Please keep full scans for personal use or study only, and avoid distributing them.
When should I use the Black & White document mode?
Use it when the paper looks yellowed or the lighting is uneven; it makes text crisper and keeps the file smaller. Keep Original for pages where color actually matters, like photos.
What does turning on OCR actually change?
It adds an invisible text layer inside the PDF so you can select and copy text, or search it with Ctrl+F. The scanned image itself stays exactly as it looked; small recognition mistakes never change the picture.
Can I scan several pages at once?
Yes. Add multiple photos and they appear as a page list you can drag to reorder, or move with the ◀ ▶ buttons. They are combined into one PDF in that order.
Are my photos uploaded to a server?
No. Corner detection, perspective correction and OCR all run inside the browser you're using right now. Your photos and the resulting PDF never leave your device.
Tools that go well with this one
JPG to PDF
Combine already-clean photos or images into one PDF.
PDF & docsCompress PDF
Shrink a scanned PDF before emailing or submitting it.
PDF & docsImage to Text (OCR)
Pull text straight out of a single photo, no scanning needed.
PDF & docsDocument submission checklist
From scanning to merging and compressing, step by step.
Guide