Document Scanner with OCR and PDF Export
Scan documents with your camera, auto-detect edges via OpenCV, extract text via OCR & export to PDF — entirely in your browser. No server. No cloud. No compromise.
Capture or upload an image to begin
No pages yet. Scan & add pages.
Fort Knox Privacy: Your Documents Never Leave Your Device
In an era of cloud-everything, Smart DocuScan ULTRA MAX takes the opposite approach. Every operation — camera capture, edge detection, perspective correction, text extraction, and PDF generation — happens exclusively in your browser using WebAssembly and Canvas API.
No file ever touches a server. No OCR data is logged. No analytics tracks your document content. Ideal for sensitive documents like contracts, passports, medical records, or financial statements.
- Zero Data Leakage: Documents stay in browser memory only.
- Truly Offline: Works with no internet after initial page load.
- WebAssembly Powered: Native-speed OpenCV & Tesseract in-browser.
- No Extension Needed: Works in any modern browser, no plugins.
Powered by Industry-Leading Open Source Technology
OpenCV.js Edge Detection
The same computer vision library used by Google, Microsoft and NASA — compiled to WebAssembly. Detects document boundaries using Canny edge detection and contour analysis, then applies a precise 4-point perspective transform to produce a perfectly flat, deskewed scan.
Tesseract.js OCR Engine
The gold-standard open-source OCR engine originally developed by HP and now maintained by Google — running entirely offline in WebAssembly. Supports 100+ languages. Extracts text from your scanned documents with impressive accuracy for printed text.
jsPDF Multi-Page Export
Compile multiple scanned pages into a professional-grade PDF document with proper page dimensions (A4, Letter, A3, Legal) and configurable JPEG quality. Generated entirely in-browser with zero server dependency — ideal for offline document archiving.
Canvas API Pro Filters
Three professional-grade image processing filters built with native Canvas API: Magic Color (auto-contrast with histogram equalization simulation), Grayscale (luminance-correct BT.709 conversion), and high-contrast B&W (adaptive thresholding for clean document reproduction).
How to Scan a Document in 5 Steps
Open Camera or Upload
Click "Start Camera" to use your device camera (mobile users get front/back switch) — or click "Upload" to load an existing photo from your device storage.
Capture the Document
Hold your camera flat over the document in good light. Click "Capture". The smart edge detector will automatically identify the document corners and place crop handles on them.
Adjust Crop Handles
If the auto-detection missed a corner, simply drag any of the 4 corner handles to fine-tune the selection. The perspective transform will flatten any angle distortion automatically.
Choose a Filter & Apply
Select Magic Color for enhanced readability, Grayscale for classic scans, or B&W for sharp printable documents. Click "Apply Crop" to process with OpenCV.
Add Pages & Export
Click "Add Page" to add the processed image to your document. Repeat for multiple pages. Then click "Export as PDF" to download your complete document — 100% offline.
Frequently Asked Questions
Does it work on mobile phones?
Yes, fully! On mobile, the tool accesses your rear camera by default (best for scanning). You can switch to the front camera using the button. All touch interactions including dragging crop handles are fully supported.
Does it need an internet connection?
Only for the initial page load to download OpenCV.js and Tesseract.js (WebAssembly files). Once loaded, the entire tool operates completely offline. Your documents are never sent anywhere.
What image formats are supported for upload?
JPG, PNG, WEBP, GIF, and HEIC (on supported browsers). The browser's native image decoder handles the format, so any format your browser supports will work seamlessly.
How many languages does OCR support?
The dropdown includes 10 common languages including English, French, German, Spanish, Chinese, Japanese, Arabic, and Hindi. Tesseract.js supports 100+ languages — additional languages can be added by a developer easily.
Initializing secure discussion portal...