Scanned PDF OCR

Upload a scanned PDF with no embedded text layer to run on-device OCR across every page and extract readable text — distinct from a regular PDF-to-text tool, which only reads text that's already embedded in the file.

OCR & Text Extraction Runs entirely in your browser

Your file is processed entirely on your device by an in-browser OCR engine — it's never uploaded to a server.

This tool is for scanned PDFs (image pages with no text layer) — each page is converted to an image and its text is extracted with OCR, which is a different process from the PDF-to-text tool that reads an existing text layer. Results are approximate and may contain errors, especially with low-quality scans.

Drag a scanned PDF here, or choose one

One file, up to 30 pages

Choose a file

What this tool does

Extracts text from a scanned PDF — one whose pages are images rather than text — by running optical character recognition over every page in a language you choose.

When to use it

A scanned document you cannot search or copy from.
An old contract or record whose text you need transcribed.
A file that arrived as images inside a PDF.
A paper archive you have scanned and want searchable.

How to use it

Put the PDF on the drop area or choose it.
Choose the "Text language in the document".
Click "Extract text from all pages" and wait for the engine to load and process.
Click "Download TXT".

A practical example

Input

A scanned 5-page contract.

Output

The text of all five pages transcribed in page order, downloadable as a text file.

Limitations

If the file already has a text layer, OCR is unnecessary — use the PDF to Text tool, which is far faster and far more accurate.
Processing works page by page, so long documents take noticeable time.
Accuracy follows the scan quality: a low-resolution or skewed scan gives weak results.
Password-protected files need unlocking first.
Formatting and tables are not recovered; the output is raw text.

Privacy and processing

All processing happens inside your browser and the file is never uploaded — essential with scanned documents, which are usually contracts and official records.

Technical notes

Each page is rendered as an image and then passed through the recognition engine, which is why this is far heavier than reading an existing text layer.