explosion/spacy

spaCy is a Python library for Natural Language Processing (NLP), the branch of AI that deals with understanding and analyzing human language in text. It provides tools for common language processing tasks: tokenization (splitting text into words and sentences), part-of-speech tagging (identifying nouns, verbs, adjectives), named entity recognition (finding people, organizations, and locations mentioned in text), dependency parsing (understanding sentence grammar structure), and text classification. It also integrates with transformer-based models like BERT, which are large neural networks pre-trained on massive amounts of text that can be fine-tuned for specific language tasks. spaCy is designed to be production-ready rather than a research tool, it prioritizes speed and reliability suitable for real-world applications that process large volumes of text. It ships with pretrained language pipelines for over 70 languages that can be downloaded and used immediately without any training. For custom needs, it includes a full training system to create your own models from labeled data. Someone would use spaCy when building applications that need to extract meaning or structure from text: a customer support system that routes tickets by topic, a legal document analyzer that finds mentioned entities and dates, a news aggregator that identifies key people and organizations in articles, or an information extraction pipeline that pulls facts from scientific papers. The tech stack is Python with Cython (a language that compiles Python-like code into fast C extensions) used internally for performance-critical parts. It installs via pip or conda and can use GPU acceleration through CUDA.