Back to Home
Free & Private OCR
Model: PaddleOCR v4On-device inference
OCR Workspace
Drop or click to upload an image
Supports PNG, JPG, WebP
Recognition Mode
Keeps a larger input size for small text and more complex scanned layouts.
Accurate Mode
PrivaKit prefers Worker OCR to keep the main UI responsive, and automatically falls back to the existing local OCR when needed.
Recognition Result
Run OCR on contracts, invoices, IDs and internal documents locally. From import to export, document content stays on your device. It is free because AI runs on your machine with no server bill; it is private because files never leave your browser.