The CockroachDB MCP Server provides a natural-language interface designed for LLMs and agentic applications to manage, monitor, and query data stored in CockroachDB. It integrates seamlessly with MCP clients, such as Claude Desktop or Cursor, enabling AI-driven workflows that interact directly with your database. The server exposes four main tool categories—Cluster Monitoring, Database Operations, Table Management, and Query Engine—each offering commands and insights to simplify database administration and analytics. Cluster Monitoring includes health checks, node status, active queries, performance statistics, and replication/distribution status. Database Operations handle connecting to CockroachDB, listing, creating, dropping, and switching databases, as well as connection status and active sessions and settings. Table Management covers creating, dropping, and describing tables and views, bulk import, index management, and schema metadata. Query Engine executes SQL with output formats (JSON, CSV, table), supports multi-statement transactions, explains query plans, and tracks query history. The MCP Server uses stdio transport by default; streamable-http support is planned. Configuration supports CLI arguments and environment variables with CLI > env > defaults. Integrations include OpenAI Agents SDK, Augment, Claude Desktop, VS Code Copilot, and Cursor. Testing with MCP Inspector.