Optical character recognition
Paper says it
Your keyboard edits it
Hand us a scan, a photograph or a PDF that is really just pictures of pages. We read the characters and hand back text you can select, search and rewrite.
Drop images or PDFs here to extract text
JPG, PNG, TIFF, BMP, WEBP, GIF, PDF, JP2 — up to 10MB
What the reader handles
Character-level reading
The engine works line by line across mixed fonts, columns and slightly skewed pages, then rebuilds the reading order.
70 languages
Latin, Cyrillic, Greek, Arabic, Hebrew and CJK scripts are all covered. Pick up to 3 at once for pages that mix them.
8 ways out
Take the raw text, drop it into a Word file or a spreadsheet, or keep the original page and gain an invisible text layer.
Page-by-page on PDFs
Multi-page documents are split first and read one page at a time, so a bad scan halfway through does not spoil the rest.
Nothing kept
Your upload is removed the moment the read finishes. We do not index, train on or retain the text we pull out.
Findable archives
Run a shelf of old scans through once and the whole set becomes searchable with your operating system's own search.
What you can feed it
Pick a source below to jump straight to the pairing you need, or just drop a file at the top and choose the output there.
Reading a whole archive?
Paid plans move your jobs onto dedicated workers with a larger per-file ceiling and a monthly minute allowance. The starter tier stays open for the occasional page.