Repair & OCR
Repair, OCR & Compare PDFs
Repair and OCR a PDF free, in your browser: rebuild a file that won’t open, turn a scan into searchable, copyable text, and compare two versions of a document word by word. Every file is read on your device; nothing is uploaded.
Tools last updated
3 ready now
Repair & OCR tools
Fix broken files, make scans searchable and compare two versions. Every one runs on your own device.
- Repair PDF Rebuild a damaged PDF that won’t open, and recover what’s readable.
- OCR PDF Make a scanned PDF searchable and copyable — the text is read on your device.
- Compare PDFs Two versions side by side, every added and removed word highlighted.
- Summarize PDF Soon The key sentences of a long PDF, picked out on your device.
- Chat with PDF Soon Ask a question, get the passages that answer it — with page numbers.
Repair & OCR guides
Common problems, fixed
Can OCR really run without uploading?
Yes. OCR here is Tesseract, an LSTM neural network compiled to WebAssembly and run in your browser tab. The engine and the language model download once, from this site; after that it works offline, and the scanned pages never leave your device.
Repair works the same way: qpdf and pdf.js, compiled for the browser, rebuild a damaged file’s index and recover every page they can read. Compare reads the text of both versions with pdf.js and matches them word by word.
Terms OCR
Questions
Why won’t my PDF open?
Usually the file was cut short by a failed download or email, or its index (the cross-reference table) is damaged. Repair PDF rebuilds the index and recovers every page it can read.
How accurate is the OCR?
On a clean, straight scan of printed text, nearly every word comes out right. Handwriting, faint copies and photos taken at an angle do worse, and OCR PDF shows a confidence score for each file so you know when to check.
Can I compare two contracts?
Yes. Compare PDFs lists every added and removed word with its page, and lays both versions side by side with the changes highlighted.