ffocr

Browser OCR with PaddleOCR + ONNX Runtime Web

Pick an image and run OCR directly in the browser. On GitHub Pages, the model URL is prefilled to the hosted PP-OCRv5 asset path.

Model Base URL

Image File or URL

Built-in samples

If no file is chosen, the URL above is used for OCR.

Model Variant Execution Provider

Min Confidence: 0.00 Lines below this confidence score will be hidden from overlay and text output.

Default sample image is ready — click Run OCR or choose another source.

Detected Text