Search

This is a search interface for a curated database of Japanese NLP resources covering everything from morphological analyzers like MeCab and Sudachi to pretrained models, datasets, and tutorials. It accepts queries in any language and converts them to English stems before matching against repository names, descriptions, categories, and tags, then ranks results by text relevance plus popularity metrics pulled from GitHub stars and Hugging Face downloads. The skill includes domain-specific logic that automatically expands queries like "sentiment analysis" to include relevant tool names, and it runs the entire search in Bash using Python to avoid hitting token limits on the JSON file. Useful when you need to find which tokenizer, BERT variant, or corpus exists for Japanese and want results sorted by actual usage rather than just keyword spam.

Install to Claude Code

npx -y skills add taishi-i/awesome-japanese-nlp-resources --skill search --agent claude-code