zemberek-nlp

Why we included this project

Turkish's agglutinative morphology is where most general-purpose NLP stacks give up, and Zemberek-NLP is one of the few open-source toolkits that handles it properly. It ships as Java libraries you can pull into a Maven build, covering morphological analysis and word generation, tokenization with sentence boundary detection, spell checking and noisy-text normalization, named entity recognition, text classification through a Java port of fastText, and language identification. Everything also bundles into a single full jar with console applications, and a gRPC server exposes the same tools to other languages, so a Python or Go service can reuse the pipeline without touching Java. One caveat: the project is in slow maintenance mode, with the last release in 2019, so treat it as a stable, proven foundation rather than something still gaining new features.