Skip to main content
Version: v1.6.47-dev1 🚧

Snowflake MCP Server

Overview​

Snowflake is a cloud-based data warehouse platform. The Snowflake MCP server enables Enterprise h2oGPTe agents to execute SQL queries, manage data warehouses, analyze data, and interact with Snowflake databases through natural language commands.

For more information, visit Snowflake.

When to Use​

This MCP server is useful for:

  • Enabling Enterprise h2oGPTe agents to execute SQL queries on Snowflake data warehouses and retrieve analytical results
  • Allowing agents to analyze large datasets stored in Snowflake and provide insights through natural language conversations
  • Integrating Snowflake data with Enterprise h2oGPTe's knowledge base for comprehensive business intelligence and reporting
  • Automating data analysis workflows where agents can query Snowflake tables, generate reports, or perform ad-hoc analysis

Requirements​

Before configuring this MCP server, ensure you have:

  • Snowflake account
  • Snowflake account identifier and username
  • A config.yaml file defining your Snowflake Cortex AI MCP services (for example, Cortex Agent, Search, and Analyst services)
  • Appropriate database and schema access

Key Features​

The Snowflake MCP server provides the following capabilities:

  • Execute SQL queries on Snowflake databases using natural language
  • Query tables, views, and schemas in Snowflake data warehouses
  • Perform data analysis and aggregations on large datasets
  • List databases, schemas, and tables in Snowflake accounts
  • Generate query results formatted for Enterprise h2oGPTe's conversational interface

Configuration​

Add the following configuration in Agents > Tools > New Tool > Remote MCP Tools:

{
  "snowflake": {
    "command": "uvx",
    "args": [
      "--from",
      "git+https://github.com/h2oai/snowflake-cortex-ai-mcp",
      "mcp-server-snowflake",
      "--service-config-file",
      "config.yaml"
    ],
    "env": {
      "SNOWFLAKE_ACCOUNT": "os.environ/SNOWFLAKE_ACCOUNT",
      "SNOWFLAKE_USER": "os.environ/SNOWFLAKE_USER"
    }
  }
}

Authentication​

Set up the following environment variables in Agents > Authentication:

  • SNOWFLAKE_ACCOUNT
  • SNOWFLAKE_USER

Notes​

  • This MCP server uses STDIO transport protocol.
  • Requires Python and the uv package manager to be available for the uvx command.
  • Requires a Snowflake account identifier and username. Ensure your user has the necessary database and warehouse access.
  • The config.yaml file controls which Snowflake Cortex AI services and tools are exposed as MCP tools. Refer to the Snowflake Cortex AI MCP documentation for full configuration details.
Feedback