Skip to content

Snowflake MCP — Data Warehouse Queries for AI Agents

What It Is

Snowflake MCP exposes Snowflake data warehouse operations over the Model Context Protocol. AI agents can run SQL queries, explore database schemas, manage virtual warehouses, and analyze enterprise data — all RBAC-aware with the user's own Snowflake permissions.

Tools Available

Tool Description
SQL query execution Run read-only or read-write queries through the agent
Schema exploration List databases, schemas, tables, views, and columns
Warehouse management Start, stop, resize, and monitor virtual warehouses
Query history Review past queries, execution times, and credits consumed

Quick Start

npx -y mcp-snowflake-server

Business Use Cases

  1. Ad-hoc analysis: "What were our top 10 customers by revenue last quarter?"
  2. Schema discovery: "Show me all tables in the sales schema with column definitions"
  3. Cost management: "Which queries consumed the most credits this week?"

Limitations

  • Snowflake account required. RBAC applies — agent can only access what the authenticated user can. Write operations need explicit enablement.

See Also