isaacwasserman/mcp-snowflake-server
FreeSnowflake 集成，实现读和（可选）写操作以及洞察跟踪。
FreeFree tier
Inputs: textOutputs: text
About isaacwasserman/mcp-snowflake-server
A Model Context Protocol (MCP) server implementation that enables database interaction with Snowflake. It supports executing SQL queries (SELECT, INSERT, UPDATE, DELETE, CREATE TABLE) through a set of tools, exposes schema context as resources, and provides a continuously updated memo for aggregating data insights. Designed for use with Claude Desktop and other MCP clients, it includes optional write capabilities and supports authentication via user/password or private key.
Key Features
read_query: Execute SELECT queries to read data
write_query (optional): Execute INSERT, UPDATE, DELETE queries with --allow-write flag
create_table (optional): Create new tables with --allow-write flag
list_databases: List all databases in Snowflake instance
list_schemas: List schemas within a database
list_tables: List tables within a database/schema
describe_table: View column definitions for a specific table
append_insight: Add data insights to the memo resource
memo://insights: Continuously updated resource aggregating discovered insights
context://table/{table_name} (if prefetch enabled): Per-table schema summaries as resources
Pros & Cons
Pros
- Open source and free to use
- Supports both read and optional write operations
- Exposes schema context as resources for deeper AI understanding
- Easy installation via Smithery or UVX
- Includes memo resource for aggregating data insights
Cons
- Requires a running Snowflake instance with proper credentials
- Write operations must be explicitly enabled via --allow-write flag, which can be a security risk if misconfigured
- No built-in authentication beyond user/password or private key (no OAuth/SAML mentioned)
- Limited to SQL-based interactions; no support for Snowflake's semi-structured data functions like FLATTEN or XML parsing
Best For
Running SQL queries on Snowflake databases from AI assistantsExploring and mapping database schemas for context-aware analysisCapturing and aggregating data insights for continuous monitoringAutomating data access and analysis workflows with MCP-enabled tools
FAQ
How do I install this MCP server?
You can install via Smithery using 'npx -y @smithery/cli install mcp_snowflake_server --client claude' or via UVX with individual configuration parameters.
How do I enable write operations?
Add the '--allow-write' flag when starting the server. This enables INSERT, UPDATE, DELETE, and CREATE TABLE queries.
What authentication methods are supported?
Authentication supports user/password via --user and --password flags, or private key via --private_key_path. You can also use a TOML configuration file.
What is the memo://insights resource?
It is a continuously updated memo resource that aggregates data insights appended via the append_insight tool. It updates automatically when new insights are added.
Can I see table schemas as resources?
Yes, if prefetch is enabled, per-table schema summaries including columns and comments are exposed as context://table/{table_name} resources.