Sign in Book a demo
Back to Tools
Databases OAuth
Build Your Agent View Snowflake Docs
snowflake logo

Snowflake MCP Tools for AI Agents

Let an agent discover accessible databases, inspect schemas and tables, read column definitions, and run read-only SELECT queries in Snowflake.

Available Tools

13 tools available in this integration

Snowflake.DiscoverDatabases

List the databases your role can access, newest-created first.

Snowflake.DiscoverSchemas

List the schemas in a database that your role can access.

Snowflake.DiscoverTables

List the tables and views in a schema that your role can query.

Snowflake.DiscoverWarehouses

List the warehouses your role can run queries on.

Snowflake.ExecuteSelectQuery

Run a read-only SELECT query and return the rows.

Snowflake.GetTableSchema

Get a table or view's columns, so you can select exact fields.

Snowflake.ListColumns

Get a table or view's columns, so you can select exact fields.

Snowflake.ListDatabases

List the databases your role can access, newest-created first.

Snowflake.ListSchemas

List the schemas in a database that your role can access.

Snowflake.ListTables

List the tables and views in a schema that your role can query.

Snowflake.ListWarehouses

List the warehouses your role can run queries on.

Snowflake.QueryRows

Run a read-only SELECT query and return the rows.

Snowflake.WhoAmI

Return the Snowflake identity every other tool uses, and the warehouse.

Why agent-optimized tools matter

Built for intent, not API schemas

Ask the agent to "find the tables containing customer churn data and compare this quarter with last quarter." Arcade picks the right Snowflake tools without guessing database, schema, table, or column names.

Up to 100x fewer tokens

Snowflake discovery reads return lightweight database, schema, table, column, warehouse, and identity metadata by default, so the agent can choose the right objects without loading query results or entire datasets into the model.

Complete the workflow end to end

Identify the account and warehouse, discover the database and schema, inspect table columns, run a read-only SELECT, and return the rows through one toolkit designed around the complete data exploration task.

Make actions predictable and controlled

Arcade tools return structured results and useful errors across users and runs. Teams can apply scope restrictions, redaction, and approval policies so agents follow the same controls every time.

Securing Snowflake

The more capabilities agents get, the more useful they are. However, more capabilities bring more risks of undesired effects such as data loss and exfiltration.

This is why Arcade provides you with contextual rules that dictate how and when a tool is allowed to run. Arcade’s contextual access lets you govern tool visibility and behavior at every stage of execution and control who can see a tool, what inputs are allowed, and what comes back.

Learn about Contextual Access Policies

For Snowflake, you can implement some of these policies:

Access when tools are listed for a user
  • Decide which tools an agent is allowed to see at all. A tool that is never listed cannot be called, whoever is asking.
Pre-execution before each tool call
  • Inspect the inputs a call is about to run with, and block it or send it for human approval when they fall outside what the agent should do unsupervised.
Post-execution after the tool returns
  • Redact Snowflake.ExecuteSelectQuery results, removing Social Security numbers, payment card numbers, email addresses, and phone numbers before they reach the model.
  • Deny Snowflake.DiscoverDatabases output when it includes databases on your restricted-data list, so an agent cannot learn the names of regulated or confidential data stores.
  • Strip Snowflake.WhoAmI results of account identifiers and role details before the model sees them, retaining only the warehouse needed to plan a query.

How to use Snowflake with your agent

Connect Snowflake to an MCP client through Arcade, then authorize each user's Snowflake account and start querying the data their role can access.

Your questions, answered

A Snowflake MCP server exposes Snowflake data discovery and read-only query capabilities as structured tools that an AI agent can call through the Model Context Protocol. After a user authorizes Snowflake, the agent can inspect accessible databases, schemas, tables, columns, warehouses, and query results permitted by that user's role.

Snowflake's own MCP server is a Snowflake-provided endpoint for accessing Snowflake capabilities. Arcade adds per-user OAuth, agent-oriented discovery and query actions, approval and redaction policies, and one MCP gateway or tool API that works across supported clients.

Arcade agents can identify the connected Snowflake user and warehouse, list accessible databases, schemas, tables, views, and warehouses, inspect table columns, and run read-only SELECT queries. They cannot insert, update, delete, create, alter, or drop Snowflake objects through this integration.

Arcade uses Snowflake OAuth for each user's account, and the user's active role must have permission to access the requested databases, schemas, tables or views, columns, and warehouse. Running a query also requires the role to be able to use the selected warehouse and read the referenced objects.

Yes. Each user authorizes their own Snowflake account through OAuth. Calls run with that user's Snowflake identity and role, so the agent can access only the databases, schemas, objects, warehouses, and rows that authorization permits.

Yes. Contextual Access policies can require human approval before a query or other Snowflake call that your organization considers consequential. The call proceeds only after the policy's conditions have been satisfied; this integration exposes read-only actions and has no write or destructive Snowflake actions.

Arcade proxies Snowflake API calls and does not persist your content as part of normal tool execution. Snowflake credentials are brokered and stored encrypted at rest, and the model never receives them.

Any framework or client that can connect to Arcade's MCP gateway or tool APIs works, including LangChain, OpenAI Agents, CrewAI, Claude Desktop, Claude Code, Cursor, Visual Studio Code, Microsoft Copilot Studio, and GitHub Copilot.

Arcade refreshes Snowflake access while the user's authorization remains valid. If it is revoked, rotated, or can no longer be refreshed, calls stop until the user reconnects the Snowflake credential.

Snowflake account, warehouse, query, and role-based resource limits still apply to calls made through Arcade. Query performance and availability depend on the warehouses and governance limits available to the authorized Snowflake role; Arcade does not remove Snowflake's limits.

Ready to actually ship your agent?

Connect your agent to Snowflake with managed OAuth, agent-optimized tools and approval controls for consequential actions.

Build Your Agent Read the Snowflake Quickstart