Rosette Entity Extraction was trained primarily on news articles. It works best with similar domains, and can be adapted to perform well on text from other genres. However, it only works on sentences; not, for example, lists or spreadsheets.
See the Rosette API documentation for more information.