Stanza

A Pythonic NLP pipeline by Stanford providing high-accuracy linguistic analysis for 70+ human languages.

Stanza delivers a suite of high-performance NLP tools built on PyTorch, including the official Python interface to the Stanford CoreNLP software. It handles everything from tokenization and multi-word token expansion to dependency parsing and named entity recognition (NER). By leveraging pre-trained neural models, Stanza achieves state-of-the-art results on the Universal Dependencies benchmarks across 70+ languages. It is the definitive choice for researchers requiring deep linguistic accuracy and developers needing a robust, multilingual pipeline for production-grade text processing.

https://stanfordnlp.github.io/stanza/