All request are made to http://api.natura.di.uminho.pt, and all responses are in JSON format. Check the online demo.
Split a text into tokens, uses FL3 by default.
Analyzes text: computes lemmas and POS tags, uses FL3 by default.
Analyzer can use NER, not used by default.
When active named entities are concaenaded with a _ and treated as a single token.
Analyze a single word, uses FL3 by default
Or Jspell can be used