This MCP server exposes a compact API for discovering datasets in the AWS Open Data Registry. It provides two main tools: search_datasets, which performs fuzzy matching against dataset names, descriptions, and tags, returning results with optional limits and detail levels; and get_dataset, which retrieves a complete dataset record given a YAML filename. The server caches the AWS Open Data Registry to speed up lookups and supports a streamable HTTP transport at /mcp, making it easy to integrate with client tools such as Claude Code. Installation and usage instructions are included, with npm install and npm start. The search_datasets operation accepts: query (optional string), limit (optional number, default 25), detail (optional enum: nameOnly, minimal, full; default minimal). The get_dataset operation requires filename (string ending with .yaml). The server is intended to be used by developers and data enthusiasts who want to programmatically discover, inspect, and retrieve metadata about AWS Open Data datasets, with simple integration via Claude Code CLI.