A table in a PDF is only text placed in rows and columns; the file does not say it is a table. Getting the numbers into a spreadsheet by copying usually runs the columns together. This looks for the tables on each page, puts each one on its own sheet, and stores numbers as numbers so they can be added up. It reads tables, not pages: text that is not in a table is not copied.
The defaults suit most files. Under Options you can choose which pages to convert and set or supply a password.
Each table found becomes a sheet. A table drawn with lines is read from its lines; one with no lines is found from the way its words line up, which is a good guess and not a promise, and some layouts defeat it. A value is stored as a number only when it can be read one way, and a code with a leading zero, a zip or a phone number stays text. Text that is not in a table is not copied, and a scanned PDF, which holds pictures of pages rather than text, ends with a plain message until OCR PDF has given it a text layer. A password-protected PDF needs its password, entered under Options; without it the file is refused.