Server is distributed through PyPI package manager for installation
Provides a Python API for query execution and catalog services
Provides SQL query execution, table previewing, metadata catalog building, and dependency graph analysis for Snowflake data warehouses with built-in safety checks to prevent destructive operations
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Igloo MCPshow me the top 10 customers by total sales this month"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Igloo MCP
A lightweight Snowflake MCP server that connects your AI assistant to Snowflake with built-in safety, caching, and auditing. Query databases, build catalogs, and create living reportsβall through natural language.
Why Igloo MCP?
π Query Safely
Block dangerous DDL/DML by default, auto-cancel slow queries, and log every execution. Use execute_query with configurable guardrails and test_connection to validate authentication before running queries.
β‘ Work Faster
Minimize token usage through progressive disclosure and smart result caching. Tools like get_report support multiple retrieval modes (summary/sections/insights/full), and search_catalog lets you find tables without hitting Snowflake.
π Stay Audited
Maintain complete query history with source attribution for compliance tracking. Every execute_query call logs to history, and Living Reports track all modifications with full audit trails via evolve_report.
π Build Living Reports
Create auditable, evolving business reports with create_report, modify them safely with evolve_report, attach charts to insights, and export to HTML/PDF/Markdown via render_report.
Quick Start
Prerequisites
Cursor Setup
Add to ~/.cursor/mcp.json:
Restart Cursor and ask: "Preview the customers table"
Claude Code Setup
Run from terminal:
Or add to ~/.claude.json manually:
Restart Claude Code and ask: "Show me the schema for my database"
Full setup guide: docs/installation.md
Core Tools
π Query & Explore
Tool | Description |
| Run SQL with guardrails, timeouts, and auto-insights |
| Export Snowflake metadata for offline search |
| Find tables/columns without querying Snowflake |
| Visualize table lineage and dependencies |
π Living Reports
Tool | Description |
| Initialize auditable JSON-backed reports |
| Modify reports with LLM assistance and audit trail |
| Perform multiple operations atomically |
| Export to HTML, PDF, or Markdown via Quarto |
| Read reports with progressive disclosure modes |
| Find reports by title or tags |
| Search citations by source type or provider |
| Discover valid structures and section templates at runtime |
π₯ Health & Diagnostics
Tool | Description |
| Validate Snowflake authentication |
| Monitor server, profile, and catalog status |
View all 15 tools: docs/api/TOOLS_INDEX.md
When to Use Igloo MCP
Choose Igloo MCP | Choose Snowflake Labs MCP |
AI assistant for dev/analytics workflows | Production Cortex AI integration |
Simple SnowCLI-based setup | Enterprise service architecture |
Query safety + automatic caching | Full Snowflake object management |
Built-in auditing and compliance | Container-based deployment |
Resources
π§ API Reference
π‘ Examples
π Changelog
π¬ Discussions
MIT Licensed | Built for agentic efficiency