Are my scanned documents or photos uploaded to any cloud server?

No. The entire Optical Character Recognition pipeline runs 100% locally in your web browser memory using WebAssembly (Tesseract.js). Zero bytes of document data or text leave your computer.

Can it extract text from multi-page scanned PDFs?

Yes. Tangent renders each PDF page to an in-memory high-density raster canvas via PDF.js, then OCRs each page sequentially, compiling the full extracted text with page demarcation.

Does it support clipboard pasting?

Yes! Simply press Ctrl + V (or Cmd + V) anywhere on the page to immediately OCR a copied screenshot from your clipboard.

100% CLIENT-SIDE PRIVACY SHIELD

Client-Side Document OCR

Extract searchable, editable text from images, photos, receipts, and multi-page scanned PDFs. Runs entirely in your browser memory via WebAssembly.

Drop your document or image here

Supports PDF, PNG, JPG, WebP, BMP, TIFF • Or press Ctrl + V to paste screenshot

Select Document

Language:

Auto-Enhance Contrast & Grayscale

Initializing WebAssembly OCR engine...

Text copied to clipboard!

WASM Neural OCR Engine

Client-side Tesseract WebAssembly recognizes printed and typed text directly on your CPU without external cloud APIs.

Zero Server Transcription

Unlike Google Cloud Vision or AWS Textract, financial receipts, contracts, and IDs never leave your computer's encrypted memory.

Formatted Text & Searchable Export

Extract clean paragraphs, copy text with one click, or export formatted .txt documents with recognized layout preserved.

Frequently Asked Questions

How accurate is the client-side OCR engine?

Our engine uses Tesseract WebAssembly trained neural models. It achieves 98%+ accuracy on clean scans, screenshots, printed documents, and receipts with clear typography.

Are scanned documents or text transmitted to an AI server?

No. All text recognition happens 100% locally within your browser sandbox. Your scanned IDs, receipts, and sensitive documents never touch an external cloud.

Which image formats and languages are supported?

The engine supports PNG, JPEG, WebP, BMP, and multi-page scanned PDFs in English, Spanish, French, German, Italian, Portuguese, Chinese (Simplified), and Japanese.