Cadmium

Composed of independent shards for tokenization, stemming, classification, etc., allowing developers to include only needed components, keeping dependencies minimal as emphasized in the installation section.

Covers a wide range from sentiment analysis and text summarization to readability scoring and language detection, providing many common NLP features directly within Crystal.

Built specifically for Crystal, ensuring type safety and high performance without relying on external services or other language ecosystems.

Includes specialized structures like tries and graphs, optimizing text processing tasks such as prefix matching and graph-based algorithms.