spaCy is an open-source Python and Cython library for production-oriented natural language processing. It provides fast, trainable pipelines for tasks such as tokenization, part-of-speech tagging, dependency parsing, named-entity recognition, text classification, sentence segmentation, and other forms of linguistic analysis.

The library supports dozens of languages and includes pretrained pipelines, transformer integration, multi-task learning, a configurable training system, model packaging, deployment tooling, and workflow management. Its APIs are designed around processing large volumes of text efficiently while still allowing developers to create custom pipeline components and train domain-specific models.

spaCy is aimed at software engineers, data scientists, and NLP practitioners building real applications rather than only experimenting in notebooks. It is commonly used for information extraction, document processing, search preprocessing, entity recognition, classification, and other language pipelines where predictable APIs, speed, reproducible training, and straightforward deployment are important.