How to Connect Snowflake to Claude / Cursor via MCP (Snowflake MCP Server Setup)

See everything Query Streams can do beyond this guide Get Started Free

Supports

+more

Live Data

Replace exports with on-demand SQL

Excel Sheets

BI & AI Analytics

Ask Nova, get SQL + charts

Meet Nova

Database REST API

One key per partner. No credentials shared.

Build an API

+more

Automation

Scheduled sync to 6+ platforms

Explore

+more

API to SQL

Query APIs with SQL, no code

Explore

+more

AI Database MCP

Claude, Cursor, ChatGPT & Grok talk to your data

Connect AI

Query Streams is a secure, real-time data integration platform that brings every database and SaaS API in your account into Claude, Cursor, ChatGPT, and Grok — through a single MCP key with no firewall changes. This guide walks through wiring up the Snowflake MCP connector specifically, so your AI tool can answer plain-English questions about your Snowflake data warehouse — multi-database analytics, customer cohorts, revenue rollups, time-travel comparisons, semi-structured VARIANT traversal — without you copy-pasting CSV exports out of Snowsight, DBeaver, or DataGrip. Learn more at QueryStreams.com and sign up for free to start asking your AI tool real Snowflake questions.

What Query Streams MCP gives you for Snowflake

Snowflake ships its own AI surfaces — Cortex Agents for in-warehouse AI workloads, plus community and partner Snowflake MCP servers that point Claude or Cursor at a Snowflake account directly. They work, but each makes a different trade. Cortex Agents keep the AI workload inside Snowflake’s compute (only Snowflake data, only Snowflake-side semantic models). Direct Snowflake MCP servers need either a publicly reachable Snowflake account or a VPN / PrivateLink path between the LLM and Snowflake. Query Streams’ Snowflake MCP server solves a wider problem: outbound-only Network Agent, multi-connector single key (Snowflake plus your Postgres, your Stripe, your HubSpot, all under one MCP config block), agent-layer read-only enforcement, and full audit trail. We’ll cover the comparison in detail further down; first, the basics.

How it works without opening firewall ports

The Query Streams Network Agent installs once on a machine that can reach your Snowflake account over outbound HTTPS to *.snowflakecomputing.com, runs SQL through the official Snowflake driver using your account’s preferred auth (key-pair RSA recommended), and dials one outbound TLS link to the cloud — nothing inbound is ever exposed, and the AI tool never sees your Snowflake account URL or your private key. how the outbound-only connection works →

The MCP key you give Cursor or Claude is scoped (read / analyze / execute), revocable at any time, and rate-limited per-key. The AI tool calls these MCP tools to do its work:

Why Schema Intelligence makes Query Streams MCP different

Most “MCP for Snowflake” servers hand your AI tool the bare INFORMATION_SCHEMA, leaving the LLM to guess what STATUS = '3' means, what EVENTS.PAYLOAD holds under its VARIANT wrapper, or whether LINE_ITEMS.ORDER_ID joins to ORDERS.ID (Snowflake foreign keys are advisory, not enforced) — which is why the first SQL against a bare warehouse schema is usually wrong. why bare schema makes LLMs guess →

Query Streams MCP returns that same schema enriched with Schema Intelligence (SI) — AI-curated metadata generated by profiling your actual Snowflake data, so every schema tool (qs_get_connector_schema, qs_get_table_schema, qs_profile_table, qs_get_relationships) returns the bare schema plus warehouse-specific hints that STATUS is a five-value enum, EVENTS.PAYLOAD is a VARIANT with discoverable colon-paths like :user_id::STRING, and time-travel queries can use AT(OFFSET => -3600). The LLM stops guessing.

To make this concrete, here is what the AI client gets back from a single qs_get_table_schema call against a typical Snowflake analytics warehouse (ANALYTICS_DB.PUBLIC.ORDERS, LINE_ITEMS, CUSTOMERS, PRODUCTS, EVENTS) — first without Schema Intelligence, then with it.

The six layers Schema Intelligence adds

Each layer addresses a class of question the LLM would otherwise guess at. The opt-in SI profiling pass runs against your Snowflake data without changing your schema — and because profiling consumes Snowflake compute credits like any read query, the agent uses a small warehouse and respects your warehouse-suspension settings. what Schema Intelligence adds and how it stays current →

Same prompt, different SQL

The proof is in the SQL the AI tool actually writes. Same Cursor session, same Claude model, same prompt — “What’s our top 10 customer revenue ranking for the last 30 days?” Without Schema Intelligence the LLM has to guess at Snowflake-specific idioms. With it, the LLM knows.

Query Streams MCP vs Snowflake Cortex Agents and Snowflake MCP servers

Snowflake ships its own AI surfaces for warehouse data. Snowflake Cortex Agents run AI workloads inside Snowflake’s compute, with semantic-model templates and direct access to your Snowflake-resident data. Direct Snowflake MCP servers (community projects, including Snowflake Labs’ own) point Claude or Cursor at a Snowflake account over the official Snowflake driver. Both work. Both are appropriate for some teams. Query Streams’ Snowflake MCP solves a different shape of problem — outbound-only, multi-connector, agent-layer enforced — and it’s worth knowing where each fits before you pick.

Both Cortex Agents and Query Streams MCP are valid answers — they coexist. Pick Snowflake Cortex Agents when the AI work happens entirely inside Snowflake’s compute on Snowflake-resident data and you want the AI workload colocated with the warehouse. Pick a direct Snowflake MCP server when you only need Claude or Cursor pointed at one Snowflake account, you’re comfortable with the network exposure, and the per-account configuration overhead is acceptable. Pick Query Streams MCP when the AI tool also needs to read your Postgres, your Stripe data, your HubSpot CRM, or your Shopify orders under one access plane — or when audit trail and agent-layer read-only enforcement are non-negotiable for compliance. Many teams deploy both Cortex Agents (for in-warehouse AI workflows) and Query Streams MCP (for cross-system AI tooling) on the same Snowflake account; they don’t overlap.

MCP not for you? Try Nova AI instead.

Skip the JSON config entirely: Nova AI is built into the Query Streams web portal and asks the same plain-English questions across your Snowflake connector plus every other connector on your account — same agent, same read-only enforcement, same Schema Intelligence, no MCP plumbing.

Meet Nova AI

Prerequisites

Before you start, make sure you have:

1. A free Query Streams account at my.querystreams.com.
2. The Query Streams Network Agent installed on a machine that can reach your Snowflake account over outbound HTTPS — see Download the Query Streams Agent. For best performance the agent should sit in the same cloud / region as your Snowflake account (e.g. an EC2 instance in the same AWS region as a Snowflake-on-AWS account).
3. A Snowflake connector configured against the agent — account URL (account.region.cloud.snowflakecomputing.com), default warehouse, and credentials. Key-pair (RSA) authentication is recommended over username/password; the agent stores the encrypted private key and the AI tool never touches it. Snowflake user role only needs SELECT on the databases you want the AI to query.
4. Any MCP-capable AI client. We’ll show Cursor, Claude Desktop, ChatGPT, and Grok in this guide; if you use Windsurf, Zed, Continue, Cline, VS Code Copilot, Codex, or Goose, the config block is essentially the same.
5. Five minutes.

Step 1: Generate an MCP key in Query Streams

Sign in to Query Streams and open the MCP page (or sign in first at my.querystreams.com and click MCP in the left navigation). Click Generate key, give the key a recognizable name (something like cursor-laptop or claude-desktop), and pick the scopes you want this key to have:

• read — the AI can browse connectors and read schema. Required for everything else.
• analyze — the AI can profile tables and discover relationships (sample values, distributions, semantic types). Strongly recommended for Snowflake work because VARIANT colon-path discovery, enum detection, and implicit FK inference all rely on it — warehouse schemas rarely declare enforced FKs.
• execute — the AI can actually run SQL. Without this, the AI is read-only against schema metadata only.

For a typical “let Claude analyse my Snowflake warehouse” workflow, all three scopes are appropriate. For a key you’re handing to a teammate or a less-trusted client, drop execute and let them browse only. You can revoke any key at any time from the same page; the AI client will see MCP_KEY_REVOKED on its next call and stop working immediately. There’s no propagation delay.

Copy the key now — Query Streams shows it once, then stores only a hash. If you lose it, generate a new one. The key looks like qsmcp_ followed by 48 random characters and is what your AI client sends in the X-MCP-Key request header.

Step 2: Add Query Streams MCP to your AI client

The configuration is the same shape across every MCP-capable client — an MCP server entry pointing at https://mcp.querystreams.com with your key in the X-MCP-Key header. Pick your client below.

Cursor ~/.cursor/mcp.json

// Edit ~/.cursor/mcp.json
{
  "mcpServers": {
    "querystreams": {
      "url": "https://mcp.querystreams.com",
      "headers": {
        "X-MCP-Key": "qsmcp_PASTE_KEY_HERE"
      }
    }
  }
}

Claude Desktop claude_desktop_config.json

// Settings → Developer → Edit Config
{
  "mcpServers": {
    "querystreams": {
      "url": "https://mcp.querystreams.com",
      "headers": {
        "X-MCP-Key": "qsmcp_PASTE_KEY_HERE"
      }
    }
  }
}

ChatGPT Apps & Connectors

// Settings → Apps & Connectors → Add MCP
Server URL  https://mcp.querystreams.com
Auth header X-MCP-Key
Header value qsmcp_PASTE_KEY_HERE

// Requires a paid ChatGPT plan
// (Plus / Pro / Team / Enterprise).

Grok Remote MCP Tools

// Grok → Settings → Tools
{
  "mcp_servers": [{
    "name": "querystreams",
    "url": "https://mcp.querystreams.com",
    "auth_header": "X-MCP-Key",
    "auth_value": "qsmcp_..."
  }]
}

Restart your AI client. On its next start it will discover the eight Query Streams MCP tools listed above and surface them in its tool palette. In Cursor and Claude Desktop you can verify by typing “list connectors” — the AI should call qs_list_connectors and return your Snowflake connector along with anything else you have configured.

Step 3: Ask the AI a Snowflake question

You don’t write SQL — the AI does. You ask a question, the AI picks the right MCP tool, the agent runs the query against your Snowflake warehouse, and the answer comes back as text plus tables. Three example prompts to try first:

The first time the AI calls a tool, your client may pop up a confirmation prompt asking you to approve the tool call — that’s MCP’s standard consent flow, not anything Query Streams adds. Approve once and the AI proceeds with the rest of the conversation freely. You can revisit the consent at any time in your client’s settings.

Frequently asked questions

Get started

Related guides: Download the Query Streams Agent | Database Connector Setup | All MCP Server guides | Nova AI text-to-SQL

Category: MCP Server

Tags: mcp, claude, cursor, snowflake, data-warehouse, enterprise, snowflake-cortex, cortex-agents, key-pair-auth, analytics

Meta Description: Connect Snowflake to Claude or Cursor via Query Streams MCP. Outbound-only, read-only, multi-connector. 5-min setup.