nlprule

Why we included this project

nlprule is a Rust library, with Python bindings, that bundles LanguageTool's rule resources into a self-contained pipeline. You get sentence segmentation, part-of-speech tagging, lemmatization, chunking, and thousands of grammar rules without standing up a model server. The API is small: load a tokenizer and rules for a language, call correct() or suggest() on text, and read back the spans, replacements, and messages. Because it works from rules and lookups rather than machine learning, it stays light on resources and fits naturally as a pre- or post-processing step in front of larger models, or as a quiet background process in an application, even client-side via WebAssembly. For teams building editors or linters, it is a practical, embeddable alternative to calling an external correction service.