...
For Web and Feed extractors, will scan the raw content of extracted documents and pull out new documents (eg from links in web pages) that are then extracted using the following subsequent pipeline elements. The original documents can either be discarded or processed with the same pipeline.
...