OCR PDF in your browser
Recognise the text in scanned pages so you can search, select and copy it. Runs on your device.
Your PDF is opened on your device. It is never uploaded.
Other PDF tools
OCR PDF in browser, without uploading the scan. The tool recognises the text on scanned pages and adds an invisible text layer, so you can search, select and copy the text in any reader. The recognition model runs on your device.
How to oCR PDF in your browser
- Choose PDF file, or drop the scanned PDF on this page. The Recognize text (OCR) dialog opens once the file is ready.
- Pick a quality: Fast (150 DPI, best for search) or Accurate (300 DPI, slower). Choose all pages, the current page, selected pages or a range.
- Press Start OCR. The first run downloads the recognition model (about 26 MB) from this site and caches it; later runs start immediately.
- When it finishes, use Search to find text, or select and copy it. Press Export to download the searchable PDF.
Why it is private
Recognition runs inside your browser: pages are rendered to images on your device and read by a PP-OCRv5 model running on ONNX Runtime with WebGPU or WebAssembly. The model files are downloaded once from this site, checksum-verified and cached; the scan itself never leaves your computer. If the main engine cannot start, an English Tesseract fallback also runs locally.
What you can do
Included
- Recognises Latin, Chinese and Japanese text with the bundled models
- Adds a real, invisible text layer that other readers can search
- Skips pages that already contain text unless you force re-recognition
- Runs offline after the model has been downloaded once
Limits
- Other scripts such as Arabic or Cyrillic need models that are not bundled yet
- Handwriting and very low-resolution scans recognise poorly; try Accurate mode
- The text layer follows recognised lines, not individual words
Frequently asked questions
Is it free?
- Yes. The PDF tools are free to use. There is no account, no payment and no watermark.
Are my scans uploaded for recognition?
- No. The OCR model is downloaded to your browser and the recognition runs there. Nothing about the document leaves your device.
Does it work offline after loading?
- Yes. After the tool and the OCR model have been downloaded once, recognition works without a connection. You can delete the cached model in the OCR panel at any time.
Which languages are supported?
- The bundled PP-OCRv5 models cover Latin-script languages, Chinese and Japanese, with automatic detection. The Tesseract fallback covers English.
Will the layout of my scan change?
- No. The page image stays exactly as it was; the recognised text is added as an invisible layer on top of it.
What is the maximum file size?
- Files up to 1 GB are accepted. Very large files depend on your device's memory, and some undo steps are skipped above 48 MB to keep the editor responsive.