OCR PDF — Scanned PDF to Text
OCR PDF extracts the text from scanned documents — PDFs that are really just photographs of pages, where normal text extraction returns nothing. Each page is rendered to an image and recognized by the Tesseract OCR engine running locally via WebAssembly, with progress shown per page. Copy the result or download it as a text file. Select several scans to queue them all: each becomes its own .txt file and the OCR engine loads once and is reused for the whole run. Sensitive scans never leave your device.
Last updated: July 2026
To OCR a scanned PDF, open it here — each page is rendered and run through the Tesseract OCR engine compiled to WebAssembly, entirely in your browser. You get the recognized text per page, ready to copy or download as a .txt file. Because nothing is uploaded, it's safe for contracts, statements and IDs.
How to use OCR PDF — Scanned PDF to Text
- Choose one or more scanned PDFs (a few pages each works best — OCR is compute-heavy).
- Pick the document's language and start recognition; progress is shown per page.
- Copy the extracted text or download it as a .txt file — or “Download all” as a .zip for a batch.
Frequently asked questions
How do I extract text from a scanned PDF for free?
Open the PDF on this page and run OCR — the recognition happens in your browser at no cost, with no page limits or sign-up. If your PDF already has selectable text, the faster PDF to Text tool reads it directly.
Can I OCR several scanned PDFs at once?
Yes. Select multiple PDFs and they're queued and recognized one at a time, each producing its own .txt file. The OCR engine is loaded once and reused across the whole queue, so a batch is faster than running the files one by one. Expect it to take a while and keep the tab open.
Why is it slow on long documents?
OCR is genuinely compute-intensive, and here it runs on your device rather than a server farm — that's the privacy trade-off. Expect a few seconds per page; very long scans are better processed in chunks (split the PDF first).
Which languages are supported?
English, Spanish, French, German, Portuguese, Italian and Hindi are offered in the picker; the underlying engine downloads the language model on first use.
How accurate is it?
On clean 300-DPI scans of printed text, accuracy is high. Skewed, low-resolution or handwritten pages recognize poorly — no OCR engine handles handwriting reliably.
Is my PDF uploaded?
No — rendering and recognition run entirely in your browser. The engine and language data download once; your document never leaves your device.
Related pdf tools
PDF to Text
Extract all text from a PDF and copy or download it. Free, private, in-browser PDF text extractor — no uploads.
Image to Text (OCR)
Extract text from an image or screenshot using free in-browser OCR. Copy the recognised text instantly — your image is never uploaded.
Split PDF
Extract pages or page ranges from a PDF into a new file — free, private and entirely in your browser. No uploads, no sign-up.
Compress PDF
Reduce PDF file size for free, right in your browser. Adjustable quality, no uploads — your file never leaves your device.