PDF to Excel

No. The PDF is read in your browser, on your own device, and is not sent to our server. Only a scanned page downloads the open-source text reader (OCR) and the language data you picked from jsDelivr; your PDF is not sent to it.

Yes. A page with fewer than 20 letters of its own text is read with OCR, and the result names the pages it read that way. Check the numbers on those pages; OCR can misread a digit.

A number becomes a real number, so you can add it up, in the convention the PDF itself uses (1.250.000,50 or 1,250,000.50); the language of this page never decides it. A signed or bracketed amount (-2.500, 1.200-, (1.200)) is negative, and $, ₫ and VND stay in the cell's format. When the PDF gives no sign of its convention (a lone 1.250), the number stays text exactly as written, and the status says so. A leading zero, a phone number, more than 15 digits, digit groups split by spaces and a tax code with a dash stay text. A date written 2024-01-05 or 5.1.2024 becomes a date cell; 05/01/2024 does when the PDF shows which number is the day (a value over 12). A percent gets a % format.

A heading that spans columns is written as a real merged cell. A cell that wraps onto several lines stays one cell, with line breaks. Two tables on one page get their own columns, one under the other with a blank row between. A page of running text, in one column or two, is not a table and is refused. A table without a gap between its columns may be read as one column, so check the preview.

Vietnamese, English, Tibetan, Chinese (Simplified and Traditional), Japanese, Korean, Thai, Khmer, Lao, Burmese, Hindi, Sanskrit, Russian, French, German and Spanish. Pick the language under Text in scans; the tool does not find it for you.

No. OCR reads printed text, typed text and clear scans. Handwriting is not supported.

It is free, with no account, because your own device does the work. One PDF can have up to 300 pages and 100 MB.