Extract text from photos and scanned PDFs with OCR
OCR can extract wording from photographs when their lettering cannot be copied. Recognition can contain mistakes, so compare the transcript with the source before sharing the extracted text.

Will OCR turn a scanned table into an editable spreadsheet?
This tool saves recognized text as TXT. It does not reconstruct spreadsheet cells or formula structure. Check names and numbers against the image, then arrange the verified text into rows and columns if you need an editable table.
How to use this tool
- Add images or a PDF and select the recognition language.
- For PDF input, choose the relevant page range.
- Run OCR and correct mistakes in the result field.
- Compare names and numbers with the source, then download TXT.
Choose the right settings
For a Korean document, select Korean + English. For PDF input, choose the page range you need so you can compare the recognized text with the source more easily.
Before you start
Define the document’s purpose
Keep the original and decide whether you need changed pages, a new document, extracted information or a report. A readable page is only part of a PDF: forms, bookmarks, annotations and existing signatures may also matter. Check this guide’s supported features before choosing the result for a formal submission.
For a tool that selects pages, distinguish positions in the file from numbers printed on the pages. For a document created from entered values, review the source values and their order instead. If size, placement or print layout can be set, test a page with dense text or content close to an edge before applying the choice more widely.
Make a short plan of the source files, their order and the changes you intend. For an existing PDF, note any page selection and the functions that must survive; for a newly generated document, keep the values from which it is built. If a problem appears only after a later operation, compare the intermediate copy with the original to locate the first change. Record the settings needed by the recipient, so a visually acceptable trial is not mistaken for a document that meets every requirement.
Check content and document behavior
Open the saved file and inspect the beginning, end and any place where content was moved, combined or divided. Compare counts and values with the source or your input list. When the output is a report or extracted text, check a few findings on the actual pages rather than treating an empty report as proof that nothing relevant exists.
If printing is the purpose, use print preview and a test sheet to check reading size and placement. Where interactive fields, links or signature verification matter, test them separately; appearance does not establish that they still work. If processing fails, review protection, format and the stated limits without discarding the original.
Example input and output
scanned page image → reviewed OCR textExamples describe the workflow. Results depend on the source file and settings.
Check the saved result before using it
Compare names, dates, amounts and account numbers with the original one item at a time. Correct recognition errors in the result field, save TXT and check the order of the lines.
If the result is different from what you expected
If a PDF exceeds the 25-page recognition limit, process selected page ranges separately. OCR recognizes printed characters; it does not reconstruct table rows, columns or equation structure.
Limits to keep in mind
- Up to 25 PDF pages are processed. Table structure and equations are not recovered; verify names, amounts and account numbers against the source.
Continue with a related tool
This guide explains the public browser tool. Keep a copy of the original and inspect the result before sharing it.