Run page-by-page GPU OCR locally with WebGPU or WASM. Recover completed pages after interruptions, download text progressively and optionally generate searchable PDF.
PDF output rasterizes selected pages as JPEG images; original vectors, forms and digital signatures are not preserved. Output includes selected pages only.
GPU support and OCR speed depend on your device. Check important outputs; complex layouts and handwriting are not guaranteed.
Waiting for a PDF.
Completed page text and (when PDF is selected) images remain in private storage on this device. Select the same PDF and settings to resume; Clear removes these files.
OCR2PDF-RAPID processes scanned PDF pages locally with PP-OCRv6 on WebGPU-capable devices and falls back to WASM when needed. It preserves existing text and produces searchable PDFs.
No. Rendering, OCR and PDF generation take place in your browser. OCR models are downloaded from Toolardo on the first run.
Yes. Automatic mode can fall back to WebAssembly (CPU).
Accuracy depends on the scan, handwriting and layout. Verify important outputs; handwriting and complex tables are not guaranteed.
