Snowflake MCP Server

An MCP server that provides read-only query functionality for Snowflake databases, supports multiple authentication methods and SQL query operations, and can be integrated with Claude for secure data access.

Database Artificial intelligence chatbots #Snowflake connection #Read-only query #MCP protocol #Data security .Python

rating : 2.5 points

downloads : 26

update time : 2025-04-29

What is Snowflake MCP Server?

The Snowflake MCP Server acts as a secure bridge between Claude AI and your Snowflake data warehouse. It allows Claude to retrieve information from your databases through controlled read-only queries while preventing any data modifications.

How to use this service?

After installation and authentication setup, you can ask Claude natural language questions about your Snowflake data. Claude will automatically translate these into appropriate database queries through the MCP server.

Ideal Use Cases

Perfect for business analysts needing quick data insights, managers requesting reports, or developers testing queries. Common scenarios include customer analytics, sales reporting, and marketing campaign analysis.

Key Features

Flexible AuthenticationSupports both automated service accounts (with private keys) and interactive browser login for different security needs.

Comprehensive Query ToolsProvides specialized tools for listing databases/views, examining structures, and executing custom queries with result formatting.

Built-in SafetyEnforces read-only operations with automatic query validation and result limiting to protect your data.

Seamless Claude IntegrationDirectly connects to Claude Desktop for natural language access to your Snowflake data.

Pros and Cons

Advantages

No SQL knowledge required - ask questions in plain English

Enterprise-grade security with Snowflake's authentication

Maintains persistent connections for faster queries

Prevents accidental data modification with read-only enforcement

Limitations

Only supports read operations (no INSERT/UPDATE/DELETE)

Requires Python 3.12+ environment

Browser authentication requires manual login sessions

Complex queries may need optimization for large datasets

Setup Guide

Install Prerequisites

Ensure you have Python 3.12+ and uv package manager installed on your system.

Clone and Configure

Clone the repository and set up your authentication method by editing the .env file.

Install Dependencies

Use uv to install all required Python packages.

Connect to Claude

Add the server configuration in Claude Desktop settings to enable integration.

Example Interactions

Database ExplorationDiscover what data is available in your Snowflake account

View AnalysisExamine the structure of specific database views

Custom QueryRequest specific business metrics through natural language

Frequently Asked Questions

Is my data secure with this connection?

Why can't I modify data through Claude?

How often do connections refresh?

Can I use this with other AI assistants?

Additional Resources

Snowflake Documentation

Official Snowflake database documentation

MCP Protocol Specs

Technical details about Model Context Protocol

Python Snowflake Connector

Reference for the underlying connection technology

🚀 Snowflake MCP Server

A Model Context Protocol (MCP) server for performing read - only operations on Snowflake databases. This tool enables Claude to securely query Snowflake data without modifying any information.

🚀 Quick Start

The Snowflake MCP Server allows Claude to query Snowflake data safely. First, you need to install the required dependencies and configure the environment variables. Then, you can run the server to start querying data.

✨ Features

Flexible Authentication: Connect to Snowflake in the following ways:
- Authenticate using a service account and a private key.
- Authenticate via an external browser for an interactive session.
Connection Pool with Automatic Background Refresh: Maintain persistent connections.
Support for Multiple Views and Databases: Query multiple views and databases within a single session.
Support for Multiple SQL Statement Types: Including SELECT, SHOW, DESCRIBE, EXPLAIN, and WITH.
MCP - Compatible Processor: Query Snowflake data.
Read - Only Operations with Security Checks: Prevent data modification.
Python 3.12+ Support: Ensure compatibility with modern Python versions.
stdio - Based MCP Server: Facilitate integration with Claude Desktop.

📦 Installation

Installation Requirements

python-dotenv
mypy
ruff
pytest
pydantic
snowflake-connector-python

Installation Steps

pip install python-dotenv mypy ruff pytest pydantic snowflake-connector-python

💻 Usage Examples

Basic Usage

Environment Variable Configuration

Create a .env file in the project root directory and add the following content:

SNOWFLAKE_AUTH_TYPE=private_key or external_browser
SNOWFLAKE_PRIVATE_KEY_PATH=Path to the private key (if using private_key authentication)
SNOWFLAKE_ACCOUNT_NAME=Account name

Running the Server

python mcp_server_snowflake/main.py

📚 Documentation

Functional Modules

mcp_server_snowflake.mcp_server: Main logic of the MCP server.
mcp_server_snowflake.models: Data models and validation.
mcp_server_snowflake.auth: Authentication handling.
mcp_server_snowflake.utils: Utility functions and configuration management.

Security Precautions

LIMIT Clause: All queries automatically include a LIMIT clause to prevent risks associated with large datasets.
Write - Operation Prohibition: Prohibit any write operations (e.g., INSERT, UPDATE, DELETE).
Input Validation: Strictly filter SQL injection risks.
Non - Persistent Credentials: Private keys and credentials are not persisted to disk.
HTTPS Communication: Use HTTPS for communication with Snowflake.

Development Information

Code Inspection Tools

# Static type checking
mypy mcp_server_snowflake/

# Code format checking
ruff check .

# Automatic code formatting
ruff format .

# Unit testing
pytest