What it does FudanNLP (FNLP) is a Java toolkit for Chinese natural language processing. It handles word segmentation, part-of-speech tagging, named entity recognition, keyword extraction, dependency parsing, and temporal phrase recognition, plus text classification and clustering. Models ship separately from source code due to size.

The interesting bit The README’s top line is a eulogy: the project was superseded by FastNLP in December 2018 and “no longer updated.” That makes this less a living tool than a historical artifact — one that still draws stars (2,690) and occasional curiosity from developers maintaining legacy Java stacks or studying how Chinese NLP was done before the PyTorch era.

Key highlights

• Pure Java implementation with LGPL 3.0 licensing
• Covers the full classic Chinese NLP pipeline: seg → POS → NER → parsing
• Ships with pre-trained models (seg.m, pos.m, dep.m) via GitHub Releases
• Published at ACL 2013, giving it academic pedigree
• Successor FastNLP is the actively maintained replacement

Caveats

• Explicitly unmaintained since 2018; no bug fixes or model updates
• Model download and manual “models” directory setup required
• Documentation and community support (QQ group, Google Code links) show their age

Verdict Worth a look if you’re maintaining a legacy Java system or studying NLP toolkit evolution. Everyone else should head straight to FastNLP.