Commerce Vector Search
Performs hybrid semantic and keyword searches across products, customers, and orders using natural language queries.
This skill empowers Claude to perform advanced data retrieval within the StateSet iCommerce ecosystem by combining vector-based semantic similarity with traditional full-text keyword matching. By leveraging OpenAI's text-embedding-3-small model and SQLite FTS5 (BM25), it enables users to find products, customer profiles, order histories, and inventory items through natural language descriptions rather than exact term matches. This is ideal for building intelligent storefronts, streamlining warehouse management, and enhancing customer support tools with highly relevant search results ranked by a combined relevance score.
主な機能
01Hybrid search combining semantic similarity and BM25 full-text keyword matching
02Comprehensive search statistics for managing embedding coverage across entities
03OpenAI-powered embeddings using the text-embedding-3-small model
041 GitHub stars
05Multi-entity support for products, customers, orders, and inventory data
06Configurable scoring thresholds to ensure result relevance
ユースケース
01Performing inventory discovery across various attributes and storage locations
02Locating specific customer records or order notes based on contextual similarity
03Finding products using vague or natural language descriptions like 'warm winter gear'