Free Online OCR (Images & PDFs)

Local AI recognition with visual bounding boxes - files never leave your device

Need to pull text out of a screenshot, photo, or scanned document? This free OCR tool runs an AI recognition model (PaddleOCR v5) directly in your browser: drop in a JPG, PNG, or PDF and get the text, with English and Chinese (including mixed) supported.

Results come with visual bounding boxes: every detected text region is outlined on the image, clicking a region jumps to its text, and you can copy lines individually or everything at once.

Unlike cloud OCR services, there is no sign-up and no usage quota - and most importantly, your files never leave your computer, so IDs, contracts, and invoices are safe to process. With WebGPU enabled, speed is close to a native app.

How to use

Click "OCR" to open the tool
Drop in an image (JPG/PNG) or a PDF file
Wait for the local AI model (downloaded once, then cached)
Review results with bounding boxes and copy the text

FAQ

Is it really free? Any limits?

Completely free with no usage limits. The model runs in your own browser, so there are no server costs to pass on.

Which languages are supported?

The current model covers English, Simplified Chinese, and Traditional Chinese, including mixed text.

Why does the first run take a moment?

The ONNX recognition model (~20 MB) is downloaded on first use and cached by the browser - subsequent runs start instantly.

Does it handle multi-page PDFs?

Yes. PDF pages are rasterized and recognized page by page, and you can browse results per page.