Snowflake

You can run a Snowflake-based MCP server that enables read-only SQL queries and data insights through a programmable interface. This server lets you execute read queries, explore databases and schemas, and continuously collect data insights into a memo resource for quick access and analysis.

How to use

Use an MCP client to connect to the Snowflake MCP Server. You can run read queries against Snowflake, discover databases, schemas, and table metadata, and fetch per-table schema details if prefetch is enabled. You’ll also be able to append data insights to a memo resource, which keeps a running record of insights discovered during analysis.

How to install

Prerequisites: You need Claude AI Desktop installed or access to a compatible MCP client that supports the MCP protocol. You also need the Cloud Snowflake connection details you plan to use.

"mcpServers": {
  "snowflake_pip": {
    "command": "uvx",
    "args": [
      "--python=3.12",
      "mcp_snowflake_server",
      "--account", "your_account",
      "--warehouse", "your_warehouse",
      "--user", "your_user",
      "--password", "your_password",
      "--role", "your_role",
      "--database", "your_database",
      "--schema", "your_schema"
    ]
  }
}

Additional setup and startup

To run the local server, install Claude AI Desktop App and the UV tool, then start the MCP as shown below.

# Install UV (if not already installed)
curl -LsSf https://astral.sh/uv/install.sh | sh

# Create a .env file with your Snowflake credentials
SNOWFLAKE_USER="xxx@your_email.com"
SNOWFLAKE_ACCOUNT="xxx"
SNOWFLAKE_ROLE="xxx"
SNOWFLAKE_DATABASE="xxx"
SNOWFLAKE_SCHEMA="xxx"
SNOWFLAKE_WAREHOUSE="xxx"
SNOWFLAKE_PASSWORD="xxx"
# Optional: SNOWFLAKE_AUTHENTICATOR="externalbrowser"

# Run the MCP server locally
uv --directory /absolute/path/to/mcp_snowflake_server run mcp_snowflake_server

Add the server to Claude Desktop configuration

Create a Claude Desktop config entry that points to your local MCP server. Use the full directory path to your MCP server and start command.

"mcpServers": {
  "snowflake_local": {
    "command": "/absolute/path/to/uv",
    "args": [
      "--directory", "/absolute/path/to/mcp_snowflake_server",
      "run", "mcp_snowflake_server"
    ]
  }
}

Notes

The server can filter out databases, schemas, or tables using exclusion patterns if you need to restrict what is exposed. You also get per-table context resources for prefetch-enabled setups. The data insights are appended to memo://insights via the append_insight tool, triggering automatic memo updates.

Available tools

read_query

Execute SELECT queries to read data from Snowflake and return results as structured objects.

list_databases

List all databases present in the Snowflake instance.

list_schemas

List all schemas within a specified database.

list_tables

List all tables within a specified database and schema.

describe_table

Describe a table to retrieve column definitions, types, nullability, defaults, and comments.

append_insight

Add a new data insight to the memo resource and trigger its update.