Execute SELECT queries to read data from the database
Execute INSERT, UPDATE, or DELETE queries (enabled only with --allow-write)
Create new tables in the database (enabled only with --allow-write)
List all databases in the Snowflake instance
List all schemas within a specific database
List all tables within a specific database and schema
View column information for a specific table
Add new data insights to the memo resource
The Snowflake Database Connector provides a seamless interface between Claude and your Snowflake data warehouse. It enables direct SQL query execution, schema exploration, and data analysis capabilities without leaving your AI assistant. With this connector, you can explore database structures, run analytical queries, and collect insights about your data. The server exposes both read and write operations (when enabled), making it suitable for both data analysis and database management tasks.
The Snowflake Database Connector allows Claude to interact directly with your Snowflake data warehouse. It provides tools for executing SQL queries, exploring database schemas, and analyzing data.
The easiest way to install is through Smithery:
npx -y @smithery/cli install mcp_snowflake_server --client claude
Add the following configuration to your Claude Desktop config file:
"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"
]
}
}
Install the UV package manager:
curl -LsSf https://astral.sh/uv/install.sh | sh
Create a .env
file with your Snowflake credentials:
SNOWFLAKE_USER="your_email@example.com"
SNOWFLAKE_ACCOUNT="your_account"
SNOWFLAKE_ROLE="your_role"
SNOWFLAKE_DATABASE="your_database"
SNOWFLAKE_SCHEMA="your_schema"
SNOWFLAKE_WAREHOUSE="your_warehouse"
SNOWFLAKE_PASSWORD="your_password"
Add the server to your Claude Desktop configuration:
"mcpServers": {
"snowflake_local": {
"command": "/path/to/uv",
"args": [
"--python=3.12",
"--directory", "/path/to/mcp_snowflake_server",
"run", "mcp_snowflake_server"
]
}
}
The server supports several configuration options:
--allow-write
: Enable write operations (disabled by default for safety)--log_dir
: Specify a directory for logs--log_level
: Set logging level (DEBUG/INFO/WARNING/ERROR/CRITICAL)--exclude_tools
: Disable specific toolsYou can also customize database/schema/table exclusion patterns by modifying the runtime_config.json
file.
Once installed, you can ask Claude to:
Query your Snowflake database: "Show me the top 10 customers by revenue from our sales database"
Explore database schema: "What tables are available in the marketing schema?"
Analyze data and collect insights: "Analyze the customer_orders table and tell me about any interesting patterns"
Create new tables (if write access enabled): "Create a new table to store customer feedback data"