Snowflake Intelligence MCP Server

Installation

1. Clone this repository

git clone [email protected]:metricinsights/snowflake-cortex-mcp-server.git

2. copy config.example.py to config.py and fill in your Snowflake details.
3. Run the service via docker

docker-compose up -d

Usage Example

import asyncio
import microcore as mc
from rich.pretty import pprint


async def main():
    mc.configure(LLM_API_TYPE=mc.ApiType.NONE)
    mcp = await mc.mcp.MCPServer("http://localhost:8009/mcp").connect()
    print("Available MCP Tools:")
    pprint(mcp.tools)

    def onprogress(progress, total, message):
        print(mc.ui.cyan(f"Progress: {progress}/{total} - {message}"))

    print("Asking MCP...")
    res = await mcp.call(
        "ask",
        agent="RETAIL_SALES",
        question="What is total sales?",
        progress_handler=onprogress,
        timeout=300,
    )
    print(mc.ui.blue(res))

asyncio.run(main())