Cloudscape Design System Documentation

This MCP server implementation provides semantic search capabilities over AWS Cloudscape Design System documentation using vector embeddings. Developed by Praveen Chamarthi, it uses the Jina Code Embeddings model and LanceDB for local vector storage to enable efficient documentation retrieval. The server exposes two complementary tools: cloudscape_search_docs for finding relevant documentation files based on natural language queries, and cloudscape_read_doc for reading the full content of specific files. The implementation features hardware-optimized embedding detection with support for Apple Silicon MPS, CUDA acceleration, and CPU fallback, along with lazy loading of resources and deduplication of search results for token efficiency in agentic workflows.