Skip to content
Data MCP Directory

SIDE BY SIDE, read 10 Nov 2026

mcp-snowflake-server (community) and Snowflake-managed MCP server, field by field: 8 and 6 tools named

mcp-snowflake-server (community)Snowflake-managed MCP server

The same fourteen fields for both servers, read 10 Nov 2026.
Fieldmcp-snowflake-server (community)Snowflake-managed MCP server
Data toolSnowflakeSnowflake
Publisherisaacwasserman (community)Snowflake
KindCommunityOfficial
Repository or docsgithub.com/isaacwasserman/mcp-snowflake-serverdocs.snowflake.com/en/user-guide/snowflake-cortex/cortex-agents-mcp
Version or commit0.4.0Docs page, MCP revision 25 Nov 2025
Last release10 Apr 2025none found on 10 Nov 2026
TransportLocal (stdio)Remote (hosted)
AuthSnowflake user credentials (password, private key, external browser) or connections.toml.Snowflake OAuth (default) or External OAuth; PAT supported; RBAC per MCP server and per tool.
AccessConfigurableConfigurable
Tools listed8 named in the evidence6 named in the evidence
Tool definitions read atraw.githubusercontent.com/isaacwasserman/mcp-snowflake-server/main/README.mddocs.snowflake.com/en/user-guide/snowflake-cortex/cortex-agents-mcp
Context weight103 tokens for 8 tool definitions (8 with a description), measured from the README readNot measured: The docs page defines user-configured tool types (CORTEX_AGENT_RUN, CODING_AGENT, CORTEX_SEARCH_SERVICE_QUERY, CORTEX_ANALYST_MESSAGE, SYSTEM_EXECUTE_SQL, GENERIC) used to build tools in a CREATE MCP SERVER specification; it publishes no fixed list of named tools, so facts.json lists tool types not tools.
StatusExists as a community server; PyPI 0.4.0 dated 10 Apr 2025, GitHub pushed 7 Oct 2025 (GitHub API).Documented by Snowflake as the Snowflake-managed MCP server (read 10 Nov 2026); the Snowflake-Labs repository points users here.
Read10 Nov 202610 Nov 2026

In one sentence each

  • mcp-snowflake-server (community): A community server for schema browsing and read-only SQL on Snowflake, with write tools off by default.
  • Snowflake-managed MCP server: Snowflake's hosted endpoint applies RBAC per server and per tool with OAuth, and serves Cortex Analyst over semantic views, Cortex Search and SQL.

The read facts, topic by topic

Each cell joins the facts read for that topic; the full list with sources is on each entry. Read 10 Nov 2026.
Topicmcp-snowflake-server (community)Snowflake-managed MCP server
Publisher and releasesThe server is a community project by isaacwasserman, not published by Snowflake; license is GPL-3.0 per GitHub API. PyPI lists mcp-snowflake-server 0.4.0 uploaded 10 Apr 2025; the GitHub repo was last pushed 7 Oct 2025.Snowflake documents a Snowflake-managed MCP server that lets agents retrieve data from Snowflake accounts without deploying separate infrastructure. The docs state support for MCP protocol revision 25 Nov 2025; no semantic version number or release date is published on the page.
Clients, auth and scopingThe server exposes a memo://insights resource and an append_insight tool that stores agent-discovered insights. When prefetch is enabled it exposes per-table schema summaries as resources.MCP clients discover and invoke Cortex Analyst, Cortex Search and Cortex Agents as tools on the managed server. Access to the MCP server does not give access to tools; privileges are granted per tool, so results depend on the calling role.
Access and safetyWrite operations are disabled by default and enabled only with --allow-write. Tool exclusion and database/schema/table exclusion patterns are supported. Authentication uses Snowflake user credentials, a private key path or external browser; the README shows plain passwords in config.Authentication is Snowflake OAuth by default, with optional External OAuth, and role-based access control on the server and tools. Docs list distinct privileges: USAGE on the MCP server plus SELECT on the semantic view, USAGE on the Cortex Search service, agent, or UDF. Docs warn that tool annotations are client hints and not security controls; RBAC is the control. The managed server is available in government regions except Azure US Gov Virginia and has not completed formal FedRAMP assessment.
What it reachesCovers Snowflake only.Serves Snowflake data only; custom tools can wrap UDFs and stored procedures.
Lineage and impactSchema tools list databases, schemas, tables and describe columns; no lineage tool is listed. None found: A lineage or dependency tool in the README tool list.Tool types are Cortex Agent, coding agent, Cortex Search, Cortex Analyst, SQL execution and generic UDF/procedure tools; none is a lineage tool. None found: A lineage, column-lineage or object-dependency tool type on the Snowflake-managed MCP server page (tool types list read).
Definitions, ownership and rulesDescribe_table returns column names, types, nullability, defaults and comments, so table and column comments reach the agent.Cortex Analyst tools use semantic views, which hold business definitions, and the server accepts semantic views only, not semantic models.
Freshness and upkeepMetadata is read live from Snowflake on each tool call; no scheduled collection is described.Tools run against live Snowflake objects; no metadata snapshot or refresh schedule is described.
ProfileRun via uvx or uv with stdio for Claude Desktop; install via Smithery is also documented.The server is a Snowflake object created with CREATE MCP SERVER, reachable at an account-scoped HTTP URL (remote).

Tool names, as published

mcp-snowflake-server (community)

read_query write_query create_table list_databases list_schemas list_tables describe_table append_insight

Snowflake-managed MCP server

CORTEX_AGENT_RUN CODING_AGENT CORTEX_SEARCH_SERVICE_QUERY CORTEX_ANALYST_MESSAGE SYSTEM_EXECUTE_SQL GENERIC

Sources

  1. raw.githubusercontent.com/isaacwasserman/mcp-snowflake-server/main/README.md, read 10 Nov 2026
  2. pypi.org/pypi/mcp-snowflake-server/json, read 10 Nov 2026
  3. docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-agents-mcp, read 10 Nov 2026

Last reviewed 10 Nov 2026.