# Snowflake MCP for AI Agents AI Agent Connect > Snowflake MCP connects your AI editor directly to your Snowflake data cloud. Run queries, map out complex schemas, and check warehouse statuses without switching tabs. It lets your agent see your tables and execute SQL in real-time. ## Overview - **Category:** industry-titans - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_FZdNhx3MFhQbRsBvKlZ0olbBDFO6A5aMdUPyNnlO/ai-agent-connect - **Tags:** sql-query, data-warehousing, cloud-data, database-schema, data-analysis ## Description Snowflake connects your AI agent to your data cloud so you can work where your code lives. Instead of jumping between your IDE and a web console, you can ask your agent to explore your data structure directly from your workspace. This Connector lets you navigate deep hierarchies of databases and schemas to find exactly what you need. You can verify that your data landed correctly in internal stages or check the status of long-running queries without hunting for IDs in a browser. It keeps your SQL accurate by allowing the agent to look at live table definitions before you write a single line of code. If you're tired of copy-pasting schemas or guessing column names, this integration puts the power of your data cloud right where you're already working. You'll find it's a lot faster to audit compute costs or verify dbt models when your agent can see the active warehouses and underlying tables instantly. It's one of the many high-quality integrations you can find in the Vinkius catalog to make your data engineering faster. ## Tools ### cancel_sql Cancel a running SQL statement ### describe_table Get table schema details ### execute_sql Executes a SQL query on Snowflake. This lets your agent perform read-only actions or test joins on the fly. ### get_session_context Get current session context ### get_statement_status Check the status of a SQL statement ### list_databases Lists all databases in your Snowflake account. This provides the high-level view needed for schema discovery. ### list_roles List security roles ### list_schemas Lists all schemas within a specific database. Use it to drill down into the specific areas of your data. ### list_tables Lists all tables within a specific schema. It helps you see the exact columns and types available for joining. ### list_users List Snowflake users ### list_warehouses Lists all virtual warehouses in your account. Use this to see what's running and avoid unnecessary costs. ## Prompt Examples **Prompt:** ``` List all the tables in the production_sales schema so I can see the columns. ``` **Response:** ``` I've checked the `production_sales` schema. Here are the tables: * `dim_customers` (Base Table) * `fct_sales_records` (Base Table) * `vw_live_revenue` (View) Would you like me to look at the specific columns for any of these? ``` **Prompt:** ``` Is my analytics warehouse running right now? ``` **Response:** ``` I checked your warehouses. The `ANALYTICS_WH` is currently **STARTED** (Size: Large). All other warehouses are currently suspended. Do you want me to run a query under this warehouse? ``` **Prompt:** ``` Run a query to find the top 5 customers by total spend. ``` **Response:** ``` I've executed that query for you. Here are the results: | Customer ID | Email | Activity Score | | :--- | :--- | :--- | | 101 | alice@example.com | 980 | | 102 | john84@test.io | 964 | | 103 | m_reyes@corp.com | 941 | Would you like me to format this as a markdown table or save the dataset? ``` ## Capabilities ### Run SQL queries in chat Ask your agent to execute SQL statements directly against your Snowflake data. ### Map database hierarchies Navigate through databases, schemas, and tables to understand your data structure. ### Monitor warehouse costs Check which virtual warehouses are active to keep your compute spend in check. ### Track query progress Get real-time status updates on long-running asynchronous queries. ### Verify data stages Confirm that files have landed correctly in your internal or external stages. ## Use Cases ### Verifying data ingestion A data engineer needs to check if a file landed in S3. They ask the agent to list internal stages and confirm the file is there. ### Building SQL joins An analyst wants to know what columns are in the sales table. They ask the agent to list schemas and tables to build a join. ### Monitoring compute costs A dev needs to see if a warehouse is running. They ask the agent to list warehouses to check for active clusters. ### Quick data exploration A user wants to run a quick count on a large table. They describe the request in plain English and the agent runs execute_sql. ## Benefits - Stop tab-switching by running execute_sql directly in your chat window. - Get accurate SQL models faster by letting your agent see list_tables definitions. - Monitor your compute spend in real-time using list_warehouses to catch active clusters. - Verify data ingestion quickly by checking internal stages with list_stages. - Track long-running jobs without guessing their status via get_query_status. - Map complex data hierarchies instantly using list_databases and list_schemas. ## How It Works The bottom line is you get a direct line from your AI agent to your Snowflake data cloud. 1. Subscribe to the Snowflake MCP and grab your Snowflake Account identifier. 2. Provide your Snowflake OAuth token or JWT key pair for authentication. 3. Ask your agent to list databases, schemas, or run a specific SQL query. ## Frequently Asked Questions **Can I use the Snowflake MCP to run queries in Cursor?** Yes, it works with any MCP-compatible client like Cursor, Claude, or Windsurf. You can run queries and check schemas directly in your editor. **Does the Snowflake MCP help with cost management?** Yes, you can use it to check which warehouses are currently active. This helps you keep a firm grip on your compute costs. **How do I connect my Snowflake account to this Connector?** You'll need your Snowflake Account identifier and an OAuth token or JWT key pair. Once provided, your agent can access your data. **Can the agent see my database schemas?** Yes, it can navigate through databases and schemas to map out your data structure. This helps you write more accurate SQL. **Does the Snowflake MCP work for dbt users?** It is excellent for dbt users who need to verify live table definitions and check data landing statuses while they are modeling. **Can I check the status of a long query?** Yes, the Connector can retrieve the status of any asynchronous query. You can monitor long-running data pipelines without refreshing your browser. **Can my AI actually read the raw table rows via an execute statement?** Yes. When the AI uses `execute_sql` with something like `SELECT * FROM schema.users LIMIT 10`, the Connector integration parses the exact row outputs. The LLM consumes the tabular data back into context so you can converse naturally about the dataset findings. **Is it completely safe to give AI power over a Data Warehouse?** Safety stems from principle of least privilege. Supply a Snowflake Token tied strictly to a read-only role or a heavily scoped down service account. This allows the AI to navigate schemas and extract data without risking destructive schema mutations like DROPs or DELETEs. **Can it search for a column name if I don't know the exact schema?** Yes! Tell your agent: 'Find which table in the SALES_DB database has a column named customer_churn_score'. Due to its autonomous workflow, the bot will pull schemas, subsequently loop over `list_tables`, query Snowflake’s internal information_schema if necessary, and deduce it entirely for you. **How do I find my Snowflake Account Identifier?** Your Account Identifier is the part of your Snowflake URL before '.snowflakecomputing.com'. It typically looks like `xy12345.us-east-2.aws`.