Create a backend which extracts all plain text words found in the input document and returns a list of them.
