Skip to content

Browser OCR ​

Drop an image into the tool below, or choose a file. Recognition, image compression and model inference all happen in this browser.

🔒 100% processed in your browser
⌁
Drop an image, or click to browse
PNG, JPEG or WebP up to 20 MB

The image is processed only in page memory and Canvas. It is never uploaded or saved.

The first run downloads about 6 MB of models. Your browser caches them for offline reuse.

Recognition result

Local recognition history

No saved recognition records yet.

Offline model cache ​

The first use downloads public Paddle OCR models (about 6 MB; exact size varies by version). The browser stores them in its HTTP cache, so recognition can continue offline while that cache is available. Clearing site data or the model cache requires another online model load.

Models are provided by ppu-paddle-ocr under its open-source license. This site does not modify or upload model files.

Browser capabilities

WebAssembly, WebGPU and memory limits vary by browser. Avoid processing several very large images at once on low-memory devices.

© 2026 Local OCR. Your images stay in your browser.