Connects Claude to PostgreSQL, Snowflake, SQL Server, BigQuery, and Oracle databases to extract complete schema metadata including tables, columns, primary keys, indexes, and constraints. Exposes two tools: extract_metadata pulls the full schema and saves it as JSON to a local directory, while query_metadata lets you search and filter the extracted data by table or column names with pagination. Useful when you need to document database schemas, analyze structure across multiple databases, or give Claude context about your data model without writing custom queries. No cloud storage required, everything writes to your local filesystem. Handles large schemas through table filtering and paginated results.
claude mcp add --transport stdio optisol-business-db-metadata-extractor-mcp -- uvx db-metadata-extractor-mcp