Snowflake

This guide walks you through connecting a Snowflake database to Roundtable so the AI can query your data using the query_snowflake tool.

Prerequisites

Before you begin, make sure you have:

• An active Snowflake account.
• A Snowflake user with access to the warehouse, database, and schema you want to query.
• Admin access to your Roundtable organization (to create connections).

tip

We recommend creating a dedicated Snowflake user for Roundtable (e.g., ROUNDTABLE_READER) with a read-only role. This keeps your personal credentials out of the system and makes access easy to audit and revoke.

---

Step 1: Gather Your Snowflake Credentials

You'll need the following values from your Snowflake account:

Field	Description	Where to Find It	Example
Account	Your Snowflake account identifier	Snowflake URL: https://<account>.snowflakecomputing.com	xy12345.us-east-1
Warehouse	The compute warehouse to use for queries	Admin → Warehouses in Snowflake	ANALYTICS_WH
Database	The database to connect to	Data → Databases in Snowflake	PRODUCTION
Schema	The default schema within the database	Inside the database in Snowflake	PUBLIC
Username	The Snowflake login username	Your Snowflake user settings	ROUNDTABLE_READER
Password	The Snowflake login password	Set during user creation	••••••••

info

The Account identifier format varies by Snowflake region and cloud provider. Common formats include xy12345.us-east-1 (AWS), xy12345.us-central1.gcp (GCP), and xy12345.east-us-2.azure (Azure). Check your Snowflake URL if you're unsure.

---

Step 2: Add the Connection in Roundtable

1. Go to your Roundtable organization's Settings → Connections.
2. Click Add Connection.
3. Select Snowflake as the connection type.
4. Fill in the required fields:

Field	Value
Connection Name	A friendly label (e.g., Production Snowflake)
Account	Your Snowflake account identifier
Warehouse	The warehouse name
Database	The database name
Schema	The schema name
Username	The Snowflake username
Password	The Snowflake password

5. Click Save.

warning

Snowflake credentials are stored in GCP Secret Manager and injected into workspace pods as environment variables. They are never exposed in the UI after initial entry. Only organization admins can view or edit connections.

---

Step 3: Test the Connection

After saving, click the Test Connection button. Roundtable will:

1. Connect to your Snowflake account using the provided credentials.
2. Verify that the warehouse, database, and schema exist.
3. Confirm that the user has query permissions.

Common test failures and their fixes:

Error	Cause	Fix
Incorrect username or password	Wrong credentials	Double-check username and password
Warehouse does not exist	Typo or warehouse is suspended	Verify the warehouse name; resume it if suspended
Database/Schema not found	Incorrect database or schema name	Check the names in the Snowflake UI
Connection timeout	Network or firewall issue	Ensure Snowflake allows inbound connections from GCP

---

Step 4: Attach to a Workspace

1. Open the workspace where you want Snowflake access.
2. Go to Settings → Connections.
3. Select your Snowflake connection from the dropdown.
4. Click Save.

The query_snowflake tool is now active in this workspace.

---

Example: Querying Your Data

Once the connection is live, workspace members can ask the AI to query Snowflake in natural language:

User:

Show me the daily active users for the past 7 days.

AI response (using query_snowflake):

The AI will:

1. Generate a SQL query like:

   SELECT
     DATE(event_timestamp) AS activity_date,
     COUNT(DISTINCT user_id) AS daily_active_users
   FROM PRODUCTION.PUBLIC.EVENTS
   WHERE event_timestamp >= DATEADD(DAY, -7, CURRENT_DATE())
   GROUP BY activity_date
   ORDER BY activity_date;

2. Execute the query against your Snowflake warehouse.
3. Return the results in a formatted table with a summary.

tip

For best results, add your table schemas and business terminology to the workspace's system prompt. This helps the AI generate accurate queries without trial and error.