Snowflake

Snowflake MCP Server enables you to run SQL queries against a Snowflake database and manage data insights through a continuously updated memo resource. It exposes a set of query, schema, and analysis tools that you can invoke from an MCP client, making it easy to read, modify, describe, and analyze Snowflake data within your workflow.

How to use

You connect to the Snowflake MCP Server from your MCP client and use the provided tools to perform common database operations. Core capabilities include reading data with read_query, inserting/updating data with write_query, creating tables with create_table, and inspecting schema with list_tables and describe-table. When you discover new insights during analysis, you can add them to the memo resource via append_insight, which also updates memo://insights automatically.

How to install

Prerequisites: ensure you have Node.js and a runtime for your MCP client, plus the ability to install and run MCP servers via your environment. You will also need access credentials for Snowflake (account, warehouse, database, user, role, and schema). Follow these steps to set up the Snowflake MCP Server for Claude Desktop:

npx -y @smithery/cli install mcp_snowflake_server --client claude

# Add the server to your claude_desktop_config.json
"mcpServers": {
  "snowflake": {
      "command": "uvx",
      "args": [
          "mcp_snowflake_server"
      ],
      "env": {
          "SNOWFLAKE_WAREHOUSE": "your_warehouse",
          "SNOWFLAKE_DATABASE": "your_database",
          "SNOWFLAKE_ACCOUNT": "your_account_identifier",
          "SNOWFLAKE_USER": "your_username",
          "SNOWFLAKE_ROLE": "your_role",
          "SNOWFLAKE_SCHEMA": "your_schema",
          "SNOWFLAKE_PASSWORD": "your_password"
      }
  }
}

Additional notes

The server exposes a single dynamic resource memo://insights that aggregates discovered insights during analysis and auto-updates as new insights are appended. Use append_insight to add insights, and consult memo://insights for the latest data-driven observations.

Available tools

read_query

Execute SELECT queries to read data from Snowflake. Input: query (string). Returns: array of objects with query results.

write_query

Execute INSERT, UPDATE, or DELETE queries. Requires --allow-write flag. Input: query (string). Returns: { affected_rows: number }.

create_table

Create new tables in Snowflake. Requires --allow-write flag. Input: query (string). Returns: Confirmation of table creation.

list_tables

Retrieve a list of all tables in the database. No input. Returns: array of table names.

describe-table

Return column definitions for a specific table. Input: table_name (string). Returns: array of column definitions with names and types.

append_insight

Add new data insights to the memo resource memo://insights. Input: insight (string). Returns: confirmation and triggers memo update.