Snowflake MCP Server

Example Use Cases

• 1

  Run arbitrary or parameterized SQL on Snowflake and receive structured row results programmatically.

• 2

  Discover and inspect Snowflake metadata (databases, schemas, tables, views, and column descriptions) to build queries or validate schemas.

• 3

  Notifications about resource changes or to monitor warehouse/query status and query history for orchestration and automated decision making.

Description

This server implements the Model Context Protocol (MCP) to expose Snowflake operations as tools — including SQL execution, metadata listing (databases, schemas, tables, views), query explanation, and warehouse status. It includes a YAML-based configuration system, configurable logging, connection/timeouts, and read-only safeguards, and supports resource subscription notifications for schema changes. Designed to run locally, in Docker, or behind corporate proxies, it is intended for use by agent frameworks and integrations needing programmatic access to Snowflake.