Stanford CoreNLP

What is Stanford CoreNLP?

Stanford CoreNLP is a Java-based natural language processing toolkit created by the Stanford NLP Group. It processes raw text through a pipeline of annotators and returns structured linguistic annotations.

Key capabilities of Stanford CoreNLP

• Tokenization and POS tagging: Splits text and labels parts of speech.
• Named entity recognition: Finds people, places, organizations and other entities.
• Dependency and constituency parsing: Produces syntactic structure.
• Coreference resolution: Links mentions of the same entity.
• Sentiment analysis: Sentence-level sentiment annotation.
• Quote attribution and relation extraction: Attributes quotes and extracts relations.
• Eight languages: Arabic, Chinese, English, French, German, Hungarian, Italian and Spanish.

How Stanford CoreNLP works

You run CoreNLP from the command line, call its Java API or start it as a web service, and choose which annotators to include in the pipeline. The text is processed in order and the result is a structured document with annotations. Recent releases require Java 11 or later.

Who uses Stanford CoreNLP?

CoreNLP is used by NLP researchers, students and engineers who need established linguistic analysis in Java. Third-party wrappers let other languages call it.

Stanford CoreNLP pricing

CoreNLP is free under the GNU General Public License v3 or later. Organizations that cannot comply with the GPL can request a commercial license through Stanford.

Stanford CoreNLP alternatives

Alternatives include spaCy, which offers fast production NLP in Python, NLTK, which is a teaching-oriented Python toolkit, and Hugging Face Transformers, which provides pretrained neural models.