Are my scanned documents uploaded to a server?

No. Scan is 100% client-side — your photos are processed entirely in your browser by OpenCV.js (edge-detect, perspective correction, enhance) and tesseract.js (OCR), then assembled into a PDF by pdf-lib. The original photos and recognized text never leave your device. The only network egress is the one-time load of the OpenCV.js and OCR engine files (cached for offline use afterwards).

How accurate is the OCR?

OCR runs on your device with Tesseract. Accuracy is good but varies with image quality, lighting, and font — a flat, well-lit, high-contrast scan recognizes far better than a skewed, dim photo. Use the B&W document mode for the cleanest text, and always review the result. We make no guarantee that OCR output is perfect.

Does it work offline?

Yes. OpenCV.js and the English OCR model are fetched once and then cached, so after the first load you can scan documents and build PDFs with no internet connection. Install Scan as a PWA and it opens and runs offline.

What if the auto edge-detection gets it wrong?

Always. You can drag the four corners to correct the selection manually, or keep the original photo unchanged. A bad auto-detect never blocks you.

Is there a watermark or a subscription?

No watermark, no subscription, and no account. This build is free. It exports clean, multi-page PDFs with optional searchable OCR text.

Your documents never leave your device.

Scan documents to PDF — right in your browser

Turn photos into clean, multi-page PDFs with auto edge-detection, perspective correction, B&W/color enhance, and on-device OCR. No upload, no watermark, no signup. Your documents never leave your device.

✓ No signup✓ No upload✓ No watermark✓ Works offline

Drop photos, paste, or choose files

One or more pages. They open locally — never uploaded.

Choose images

Auto edge-detect + flatten

OpenCV.js finds the document edges and applies 4-corner perspective correction — a skewed photo becomes a flat, straight page.

B&W doc & color modes

Adaptive-threshold B&W mode for crisp text documents, or a color-preserving enhance for photos and receipts.

Manual 4-corner fallback

Drag the corners to correct a miss, or keep the original. A bad auto-detect never blocks you.

Multi-page

Add, reorder, and delete pages. Each page is independently corrected and enhanced; thumbnails keep you oriented.

On-device OCR

Optional Tesseract OCR adds a searchable, selectable text layer to the PDF — accuracy varies, so you can review it.

Private & offline

Nothing is uploaded. Install it as a PWA and the whole scanner works offline after the engines load once.

How to scan a document to PDF

1
Drop, paste, or photograph a document (receipt, page, contract, whiteboard, ID). Everything runs in your browser — nothing is uploaded.
2
Scan auto-detects the document edges and applies perspective correction + a clean enhance. Pick B&W document mode for text or color/photo mode for images.
3
If the auto-detect misses, drag the four corners to fix it — or keep the original. Add more pages; reorder or delete as needed.
4
Optionally run on-device OCR, then download a clean multi-page PDF (with a searchable text layer when OCR ran). Your documents never leave your device.

A real scanner, fully private

Other scanners upload your documents or stamp a watermark. Scan does neither — and it's free. Scroll up to scan a page.

Frequently asked questions

Are my scanned documents uploaded to a server?: No. Scan is 100% client-side — your photos are processed entirely in your browser by OpenCV.js and tesseract.js, then assembled into a PDF by pdf-lib. The original photos and recognized text never leave your device. The only network egress is the one-time load of the OpenCV.js and OCR engine files, cached for offline use afterwards.
How accurate is the OCR?: OCR runs on your device with Tesseract. Accuracy is good but varies with image quality, lighting, and font — a flat, well-lit, high-contrast scan recognizes far better than a skewed, dim photo. Use the B&W document mode for the cleanest text, and always review the result. We make no guarantee that OCR output is perfect.
Does it work offline?: Yes. OpenCV.js and the English OCR model are fetched once and then cached, so after the first load you can scan documents and build PDFs with no internet connection. Install Scan as a PWA and it opens and runs offline.
What if the auto edge-detection gets it wrong?: You can drag the four corners to correct the selection manually, or keep the original photo unchanged. A bad auto-detect never blocks you.
Is there a watermark or a subscription?: No watermark, no subscription, and no account. This build is free. It exports clean, multi-page PDFs with optional searchable OCR text.