A PDF text layer stores the document's characters separately from its visual appearance. It makes words searchable, selectable and copyable. The characters may come from the original document export or from recognition performed on a scan.
Checking whether text is present
Select a line in a PDF viewer, copy it into a text editor and search for a word visible on the page. Selection alone is not enough: the copied text should match the visible words. Check several pages because a single PDF can contain both text and scanned images.
Where the text layer comes from
A PDF exported from a word processor often contains the original characters. A scanned document initially contains a page image; OCR can recognise its text. Recognised text can still contain incorrect characters.
Extracting text or using recognition
If copying produces the correct words, try an available text conversion. If the layer is missing, contains meaningless characters or omits pages you need, check the recognition options. Do not choose OCR solely because a file has the PDF extension: first inspect its contents.
Limits of a text layer
Copying order may differ from reading order, particularly on pages with multiple columns. A text layer does not reliably describe table rows, merged cells or original formulas. Use an appropriate table conversion and compare the output with the source; check numbers and signs separately when the data matters.