Overview
Friday can connect to your Snowflake data warehouse to run analytical queries directly on your data. You can explore table schemas, run SQL queries, identify trends and anomalies, and generate summary reports — all without downloading data locally. To enable this capability you must connect the Snowflake integration by providing your warehouse credentials. After setup, Friday will use the Snowflake integration automatically whenever a task requires it.What Friday can do with Snowflake
- Discover table schemas — describe any table to see its columns, types, and structure.
- Run analytical SQL queries — execute SELECT queries directly on your Snowflake warehouse.
- Identify trends and anomalies — analyze datasets for patterns, outliers, and statistical summaries.
- Generate analysis reports — produce summary artifacts with key metrics and insights saved to your Library.
MY_DB.MY_SCHEMA.MY_TABLE) to specify which table to analyze.
Prerequisites
- A Snowflake account with access to the data you want to analyze.
- The following Snowflake credentials:
- Account identifier (for example,
xy12345.us-east-1) - Username and password
- Warehouse name (for example,
COMPUTE_WH) - Role (for example,
ANALYSTorACCOUNTADMIN)
- Account identifier (for example,
- Optionally, a default database and schema to narrow the connection context.
Step-by-step guide
Friday will automatically detect when you need to connect to Snowflake as part of your regular use.Start with a real Snowflake request
Tell Friday what you want to analyze. Include the fully qualified table name so Friday knows where to look.Example prompts:
- “Analyze the ANALYTICS_DB.PUBLIC.DAILY_REVENUE table for trends over the last 90 days.”
- “Describe the schema of PROD_DB.REPORTING.USER_EVENTS in Snowflake.”
- “Find anomalies in the MARKETING_DB.PUBLIC.CAMPAIGN_PERFORMANCE table.”
- “Every Monday, analyze ANALYTICS_DB.PUBLIC.WEEKLY_METRICS and email me a summary.”
Connect your Snowflake credentials
If you haven’t connected Snowflake previously, Friday will prompt you to enter your credentials.
- Click Connect Snowflake.
- Enter your account identifier, username, password, warehouse, and role.
- Optionally enter a default database and schema.
- Friday validates the connection and stores your credentials securely.
Continue your task
After authentication, the task you set up in Step 1 will continue automatically, and Friday will connect to your warehouse and run the analysis.
Limitations
- Read-only — Friday can only run SELECT queries. INSERT, UPDATE, and DELETE statements are blocked.
- One table at a time — each analysis request targets a single table. To compare across tables, run separate requests.
- Server-side execution — all queries run on your Snowflake warehouse and consume your Snowflake compute credits.
- Fully qualified names recommended — use the
DATABASE.SCHEMA.TABLEformat to avoid ambiguity. If you set a default database and schema during setup, you can use shorter names within that context.
Security and permissions
- Your Snowflake credentials are stored securely and encrypted; the raw values are never exposed after setup.
- Friday connects using the role you specify. The queries Friday can run are limited to whatever that role is authorized to do in Snowflake.
- Rotating your Snowflake password disables the connection; you’ll need to reconnect with the new credentials.
Revoking access
Note: By revoking access, any current and future Spaces set up with Friday will no longer be able to use Snowflake until you reconnect. To remove the Snowflake credentials from Friday:- In Friday, go to Settings → Integrations.
- Find the Snowflake integration and click Remove.
- Change or disable the password for the Snowflake user account used with Friday.
- Alternatively, revoke the role’s privileges on the tables you no longer want Friday to access.