This is the direct connector. If you’d like to extract data (to join data across multiple sources), use the managed Snowflake connector instead.
Connecting
You’ll create a dedicated user in Snowflake, grant it access to the warehouse and data you want Evidence to read, then enter the credentials in Evidence.Prerequisites
- A Snowflake account with a warehouse and at least one database you want Evidence to query.
- A user with
SECURITYADMINandSYSADMIN(orACCOUNTADMIN) to run the setup SQL.
1. Create a role and user for Evidence
If you already have a read-only user suitable for BI workloads, skip to step 2. Otherwise, run the following, replacing<warehouse>, <database>, <schema>, and <strong-password> with your own values:
SQL to create user
SQL to create user
PASSWORD = '<strong-password>' line with RSA_PUBLIC_KEY = '<your-public-key>'. See the Snowflake docs on key-pair authentication for how to generate the key.2. Configure the connector in Evidence
Open the connectors page
Go to Connectors in the sidebar. In the Warehouse card at the top, select Snowflake.
Configuration reference
- Studio
- CLI (connection.yaml)
Credentials
Snowflake account identifier, e.g. xy12345.us-east-1.
Snowflake login name.
Password for password auth.
PEM-encoded RSA private key for key-pair auth.
Passphrase for the private key, if it is encrypted.
Compute warehouse to use for queries.
Default database.
Role applied per session.
Visibility
Production schema, plus optional developer schemas that act as preview environments.
Row-level security
Session variables
{ snowflakeVariable: string; evidenceVariable: `user.email` | `user.id` | `user.name` | `organization.id` }[]
Evidence identity → Snowflake session variable, ALTERed per query for RLS. Default:
[].
