Snowflake MCP Server for Claude
Snowflake-managed MCP server for connecting Claude and other MCP clients to Cortex Analyst, Cortex Search, Cortex Agents, SQL execution, UDFs, stored procedures, and governed warehouse data.
Open the source and read safety notes before installing.
Safety notes
- Snowflake recommends verifying third-party MCP servers and tool descriptions before use because overlapping MCP servers can create tool poisoning or tool shadowing risk.
- Prefer OAuth. Hardcoded tokens and broad Programmatic Access Tokens can leak privileged Snowflake access into client config, logs, shell history, support bundles, or AI transcripts.
- Configure the Snowflake MCP server and each tool with least privilege. Access to the MCP server object does not automatically mean the user should have access to every Cortex, SQL, UDF, or stored procedure tool behind it.
- SQL execution tools can query Snowflake data and, if configured with write-capable settings, can run mutating SQL. Review generated SQL and use read-only configuration where Claude should only inspect data.
- Avoid recursive agent or MCP configurations. Snowflake documents a maximum recursion depth for circular MCP and Cortex Agent invocations, but loops can still create cost, latency, and confusing automation behavior before they stop.
- UDF and stored procedure tools can execute business logic in Snowflake, so do not expose procedures that perform billing, deletion, notification, privilege, or workflow actions without human approval.
Privacy notes
- Tool results can expose Snowflake account identifiers, database names, schema names, table names, semantic views, Cortex Search results, SQL text, query results, citations, reasoning traces, tool calls, UDF inputs, stored procedure outputs, and warehouse metadata.
- Cortex Agent responses intentionally include intermediate steps such as reasoning traces, tool calls, search results, and citations, which can make MCP payloads large and data-rich.
- Claude transcripts, MCP client logs, terminal scrollback, screenshots, support exports, and generated summaries can retain Snowflake data outside Snowflake's normal access, retention, and audit controls.
- OAuth sessions use the connecting user's default role for Snowflake MCP access. Confirm default role and default warehouse settings before relying on per-session role selection in clients such as Claude.
- Warehouse rows, semantic search results, UDF outputs, and stored procedure results can contain prompt-injection text, secrets, customer data, financial data, product telemetry, or regulated data that should not be pasted into unrestricted AI conversations.
Prerequisites
- Snowflake account in a supported region, with the managed MCP server feature available for the account.
- Database and schema where the MCP server object will be created.
- Snowflake role with least-privilege access to the MCP server object and each exposed tool.
- Default role and default warehouse configured for each user who will authenticate through an MCP client.
- Snowflake OAuth security integration for the MCP client, or an approved Programmatic Access Token fallback when OAuth is not available.
- Reviewed Cortex Analyst, Cortex Search, Cortex Agent, SQL execution, UDF, or stored procedure tools that should be exposed to Claude.
- MCP client that supports HTTP transport and Snowflake authentication, such as Claude, VS Code, or another compatible client.
Schema details
- Install type
- cli
- Troubleshooting
- No
- Scope
- Source repo
- Estimated setup
- 30 minutes
- Difficulty
- advanced
Full copyable content
{
"mcpServers": {
"snowflake": {
"type": "http",
"url": "https://ACCOUNT.snowflakecomputing.com/api/v2/databases/DATABASE/schemas/SCHEMA/mcp-servers/SERVER_NAME"
}
}
}About this resource
Content
The Snowflake-managed MCP server lets Claude and other MCP-capable clients discover and call governed Snowflake tools without deploying a separate server beside the warehouse. Snowflake administrators create an MCP server object in a database and schema, attach selected tools, configure OAuth or token-based access, and grant roles access to the server and the underlying tools.
This is an enterprise data-access integration, not a toy local connector. Treat the MCP server as a governed interface to Snowflake data, Cortex AI features, SQL execution, UDFs, stored procedures, and agent workflows. Start with narrow read-only tools, explicit role grants, and non-production data before allowing Claude to query broader warehouses or trigger custom tools.
Features
- Snowflake-managed MCP endpoint for database and schema scoped MCP server objects.
- Snowflake OAuth based authentication, with Programmatic Access Token fallback when OAuth is not available.
- Role-based access control for MCP server discovery and individual tool use.
- Cortex Search Service query tools for unstructured search over Snowflake data.
- Cortex Analyst tools for natural-language analysis over supported semantic views.
- Cortex Agent tools for invoking configured Snowflake Cortex Agents through MCP.
- SQL execution tools, with documented read-only configuration and query timeout options.
- Custom tools backed by Snowflake UDFs and stored procedures.
- Support for standard MCP
tools/listandtools/callrequests. - Official Snowflake-Labs open-source MCP repository for Snowflake MCP workflows that need a Snowflake-backed server implementation outside the managed server path.
Use Cases
- Let Claude ask governed questions over Snowflake semantic views through Cortex Analyst.
- Search enterprise knowledge or product data through Cortex Search without handing Claude unrestricted database access.
- Expose a narrowly scoped read-only SQL execution tool for bounded warehouse analysis.
- Call approved Snowflake UDFs or stored procedures as MCP tools after human review of the business action.
- Connect an MCP client to a Cortex Agent that already has curated resources, tools, and response behavior.
- Prototype Snowflake-backed MCP workflows locally with the Snowflake-Labs MCP repository, then move production access behind managed Snowflake governance.
Installation
Snowflake setup
- Choose the database and schema where the MCP server object will live.
- Create or select the Cortex Analyst, Cortex Search, Cortex Agent, SQL, UDF, or stored procedure tools that should be exposed.
- Create the Snowflake MCP server object with a reviewed specification.
- Configure Snowflake OAuth for the client or approve a least-privilege Programmatic Access Token fallback.
- Grant the connecting role access to the MCP server and only the tools it should discover and invoke.
- Set each MCP user's default role and default warehouse to the intended least-privilege values.
Claude Code
After the Snowflake MCP server object and authentication are configured, add the server URL to Claude Code:
claude mcp add --transport http snowflake https://ACCOUNT.snowflakecomputing.com/api/v2/databases/DATABASE/schemas/SCHEMA/mcp-servers/SERVER_NAME
Start with a prompt that lists tools and summarizes what each one can access before asking Claude to query or invoke anything.
Configuration
{
"mcpServers": {
"snowflake": {
"type": "http",
"url": "https://ACCOUNT.snowflakecomputing.com/api/v2/databases/DATABASE/schemas/SCHEMA/mcp-servers/SERVER_NAME"
}
}
}
Examples
Inspect available tools first
List the Snowflake MCP tools available on this server and explain what each one can access before calling any tool.
Query through Cortex Analyst
Use the approved Cortex Analyst tool to answer this revenue question from the semantic view, then show the generated SQL before summarizing.
Search governed data
Use the product Cortex Search tool to find the top matches for this customer question and include citations from the returned search results.
Run bounded read-only SQL
Use the read-only Snowflake SQL tool to return the top 10 rows needed for this analysis, avoiding broad scans or writes.
Call a custom tool after review
Show the exact stored procedure tool call you want to make and wait for approval before invoking it.
Source notes
- Snowflake's official documentation describes the Snowflake-managed MCP server as a generally available feature for letting AI agents retrieve data from Snowflake accounts without deploying separate infrastructure.
- The documentation says MCP servers on Snowflake can expose Cortex Analyst, Cortex Search, Cortex Agents, SQL execution, UDFs, and stored procedures as tools through a standards-based interface.
- Snowflake documents OAuth-based authentication, role-based access control, the importance of least privilege, default role behavior for OAuth sessions, and default warehouse requirements.
- Snowflake's security recommendations warn about third-party MCP server verification, hardcoded token leakage, least-privilege PATs, tool-level permissions, and recursive MCP or Cortex Agent loops.
- The docs show the MCP server endpoint shape under
/api/v2/databases/<database>/schemas/<schema>/mcp-servers/<name>and describetools/listandtools/callinteractions. - The GitHub repository
Snowflake-Labs/mcpis Apache-2.0 licensed and describes an MCP server for Snowflake including Cortex AI, object management, SQL orchestration, semantic view consumption, and more.
Duplicate check
Checked current content/mcp/, content/tools/, guides, skills, agents, open
pull requests, live issue state, and repository-wide content for Snowflake MCP, Snowflake-managed MCP, Cortex Agents MCP, Cortex Analyst MCP,
Cortex Search MCP, Snowflake-Labs/mcp, SYSTEM_EXECUTE_SQL,
CORTEX_AGENT_RUN, CORTEX_SEARCH_SERVICE_QUERY, CORTEX_ANALYST_MESSAGE,
snowflakecomputing.com/api/v2, and Snowflake SQL MCP. Existing BigQuery and
database MCP entries cover adjacent warehouse access, and the TruLens entry
mentions Snowflake as the current steward of TruLens, but no dedicated
Snowflake MCP entry, Snowflake MCP source URL duplicate, or open duplicate PR
was found.
Disclosure
Editorial listing. No paid placement or affiliate link is used.
Source citations
Signals
Loading live community signals…
A short, calm digest of reviewed Claude resources. Unsubscribe any time.