Extract text from images & PDFs

Free, private, browser-based OCR. Runs 100% on your device — no uploads, no login, no fees.

100% browser-based 50+ languages Images & PDFs No uploads

Input

Drop an image or PDF, paste from clipboard, or browse.

📄

PNG, JPG, WebP, BMP or scanned PDF

Custom parameters

Runtime backend

Det threshold

Box threshold

Unclip ratio

Rec threshold

No file selected

Build a multi-page PDF from images, or a searchable PDF from an input PDF.

Everything runs locally in your browser. Your files never leave your device.

No uploads, no servers. Processing happens on your device.

ONNX Runtime Web with WebGPU/WASM acceleration.

One unified model — Chinese, English, Japanese, Arabic, and more.

JPG, PNG, WebP, BMP and scanned PDFs.

A single unified model handles all of them — no model switching.

Is it free?

Yes, completely free. No account, no API key, no quotas.

Are my files uploaded?

No. OCR runs entirely in your browser. Files never leave your device.

Which OCR engine is used?

A modern OCR model running locally via ONNX Runtime Web, with WebGPU/WASM acceleration.

Does it work offline?

After the model is cached by your browser, yes.

Mobile support?

Limited — WebAssembly memory may be constrained on some mobile devices.